Impact AI

What if AI could decode the mysteries of your microbiome for a healthier you? In this episode, I sit down with Leo Grady, Founder and CEO of Jona, to discuss his groundbreaking work in microbiome research. Jona is a health technology company that specializes in microbiome profiling and analysis. It offers microbiome testing kits for individuals to use at home, along with AI-powered analysis of the associated microbiome data. In our conversation, we delve into the human microbiome and how Jona is harnessing the power of AI to unlock its secrets and revolutionize healthcare practices. Discover how Jona bridges the gap between research and clinical practice and utilizes deep shotgun metagenomic sequencing. We discuss why he thinks AI is a critical technology for decoding the microbiome, how Jona is able to connect research findings to microbiome profiles, and the company’s approach to model validation. Gain insights into the evolving landscape of AI in healthcare, the number one barrier to clinical translation and adoption of AI technology, what needs to be done to overcome it, and much more.Key Points:Background about Leo and what motivated him to start Jona.He explains the complexity of the microbiome and its role in human health.Hear more about Jona and how the company leverages AI for data analysis.How Jona applies models to analyze microbiome data and medical literature.The technical nuances and validation processes behind the company’s approach.Learn about the challenges of building models to elucidate microbiome data.Explore the intricacies of validating the company’s groundbreaking technology.Advancements in AI and machine learning that he is most excited about.Leo shares advice for leaders of AI-powered startups.Uncover the number one barrier to AI adoption: payment. What the future looks like for Jona and what the company aims to achieve.Quotes:“What's really remarkable to me about the microbiome is that it's been linked to almost every aspect of human health.” — Leo Grady“There are a lot of challenges that forced us to really develop new kinds of [machine learning] techniques that are really suited to this problem. We can't just rely on taking what's out there today.” — Leo Grady“The AI is doing that extraction. We have human oversight to make corrections to it. But once that paper has been extracted correctly, then we don't need to look at it again. It’s a one-time review process on every study.” — Leo Grady“I think the biggest challenges with AI and healthcare today are no longer technical, and they're no longer regulatory. The fact is that with current AI technology and enough data, we can solve almost any AI problem that we want to.” — Leo GradyLinks:Leo Grady on LinkedInJonaResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Bringing transparency and accuracy to the marketplace by producing high-quality data on all types of hard problems is a main focus for today’s guest and the company he works for. I am pleased to welcome David Marvin to Impact AI. David was the Co-Founder and CEO of Salo Sciences, which was acquired by Planet last year, and is now the Product Lead for Forest Ecosystems there! He joins me today to talk about monitoring forests. We delve into his background and path to Salo Sciences and their eventual acquisition by Planet; including the original mission and vision and what they worked to accomplish at Salo. David then explains his goals and focus at Planet, and unpacks the types of satellite imagery, models, and sensors they incorporate into their data and outputs. He highlights their approach to validation, how they are reducing bias, and how they are integrating extensive knowledge to empower their machine learning developers to create powerful models.Key Points:David shares details about his background and path to Salo Sciences and Planet.The original vision and mission of Salo Sciences and what they did there.He explains how they leveraged large-scale airborne LiDAR collections and deep learning to create maps of vegetation fuels.His goals and focus at Planet.David unpacks the types of satellite imagery, models, and sensors they incorporate into their data and outputs.How they validate that their models work in places where they do not have Airborne LiDAR.Reducing the bias that results from only having data in a heterogeneous distribution of LiDAR sites around the world.How they integrate their extensive knowledge to empower their machine-learning developers in creating powerful models.The business benefits he’s seen from publishing and making it a priority.His advice to other leaders of AI-powered startups.His thoughts on the impact of the forest monitoring efforts at Planet in three to five years.Quotes:“A company like Planet was essentially probably the only company we would have really ever been acquired by just given their vision and the fact that they have their own satellites and we’re a satellite software company.” — David Marvin“[At Salo Sciences] we leveraged high-quality airborne LiDAR measurements of forests all over California. Airborne LiDAR is one of these technologies, these sensors, that was on that airplane back in my post-doc lab. It shoots out hundreds of thousands of pulses of laser light per second and reflects back to the sensor, and it can basically recreate in three dimensions a forest, or a city, whatever your mapping target is. It's extremely precise. It's centimeter-level accuracy, and it's very high-quality data. We consider that the gold standard of forest measurement.” — David Marvin“Ultimately, we want to produce a near-tree-level map of the world's forests, and we're well on our way to doing that and expect to be releasing that later this summer, or in the fall.” — David Marvin“We approach the validation aspect from a few different angles, trying to source as many different independent data sets as possible to do validation. Then we also like to do comparisons to well-known public data sets; either from academia or from governments.” — David Marvin“You really do have to have the three legs of the stool to be able to build a quality operational product that is meant for forest monitoring.” — David Marvin“Making sure you have scientists on your team, making sure you're still active in the scientific publishing community, that you're up on the latest papers that are coming out, and basically acting like a scientist in an industry position is crucial to make any product work; especially in branding markets, like forest monitoring and carbon markets.” — David MarvinLinks:David MarvinDavid Marvin on LinkedInDavid Marvin on xSalo SciencesPlanetResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Foundation models have been at the forefront of AI discussions for a while now and joining me today on Impact AI is a leader in the creation of foundation models for pathology, Senior Vice President of Technology at Paige AI, Razik Yousfi. Tuning in, you’ll hear all about Razik’s incredible background leading him to Paige, what the company does and how it’s revolutionizing cancer care, and the role machine learning plays in pathology. Razik goes on to explain what foundation models are, why they are so helpful, how to train one, the differences in training one for pathology specifically, and how they use foundation models at Paige AI. We then delve into the challenges associated with the creation of foundation models before my guest shares some advice for leaders in machine learning. Finally, Razik tells us where he sees Paige AI in the next few years.Key Points:Introducing today’s guest, Razik Yousfi.An overview of Razik’s background and what led him to become Senior Vice President of Technology at Paige AI. What Paige does and why it’s important for cancer care. The role machine learning plays in pathology. Razik tells us what a foundation model is, why it’s useful, and what it takes to train one. The subtle differences in training a foundation model for pathology versus other data. How they are using foundation models at Paige AI. Razik discusses what the future of foundation models for pathology looks like. Why Razik doesn’t suggest that every organization build a foundation model. Our guest shares some advice for leaders of machine learning teams. Where he sees the impact of Paige AI in the next three to five years. Quotes:“Paige is focusing on digital and computational pathology. In other words, we really bring AI and novel AI solutions to the field of pathology to help pathologists make better-informed decisions.” — Razik Yousfi“A foundational model is a model trained on a very large set of data. The idea there is that you can, in turn, use that foundation model to build a wide range of downstream applications.” — Razik Yousfi“Building a foundation model is not easy. So, I wouldn't necessarily recommend to every organization to build a foundation model.” — Razik YousfiLinks:Razik Yousfi on LinkedInRazik Yousfi Email AddressRazik Yousfi on XPaige AIResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Biodiversity is not just an ecological concern. As you’ll learn in this episode, it has tangible economic implications too. Today on Impact AI, I'm joined by Dr. Noelia Jimenez Martinez, Head of Insights and Machine Learning at NatureMetrics, to talk about biodiversity monitoring. NatureMetrics is a global nature intelligence technology company providing end-to-end nature monitoring and impact reporting. Powered by eDNA, their Nature Intelligence Platform allows any company to manage its impacts and dependencies on biodiversity at scale, translating the complexities of nature into simple insights that help to inform the best decisions for both the planet and business. Tuning in, you’ll learn about the importance of NatureMetrics’ work, the role that machine learning plays in their technology, and some of the challenges that come with working with sometimes unpredictable data from nature. In my conversation with Noelia, we also touched on why biodiversity is an increasingly urgent imperative for businesses of all kinds, how NatureMetrics is democratizing biodiversity monitoring, and much more!Key Points:Insight into Noelia's background in astrophysics and how it led her to NatureMetrics.What NatureMetrics does, what eDNA is, and why it’s so important for sustainability.The major role that machine learning plays in NatureMetrics' technology.Specific examples of the types of models that NatureMetrics trains.How Jurassic Park helps us understand what eDNA data looks like.Different ways that this data is gathered depending on the relevant project.Unique challenges of sampling for eDNA and training models based on those datasets.How NatureMetrics measures the impact of its technology and makes biodiversity monitoring more accessible and achievable.Noelia’s urgent and common sense advice for other leaders of AI-powered startups.What the future holds for NatureMetrics and how their impact will continue to grow.Quotes:“I couldn't focus too much on solving galaxy formation with the amount of bad news I was seeing in the climate space and biodiversity collapse. I made a transition – [to] looking for jobs to apply [my astrophysics skills to] related problems in climate and biodiversity.” — Noelia Jiménez Martínez“Nature does not seem to behave [as well] as we would want. It might be that you have exactly the same covariates and your model is predicting species, and then you go, and it's not there.” — Noelia Jiménez Martínez“[Most companies] will have to report on their sustainability strategies in the world to keep on functioning. In that context, what we can do here is make biodiversity monitoring achievable and democratically easy to access.” — Noelia Jiménez Martínez“The success of [an AI startup is] – tied up to the diverse, strong teams you build.” — Noelia Jiménez MartínezLinks:NatureMetricsDr. Noelia Jiménez Martínez on LinkedInResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

In today’s episode, I am joined by Simon Arkell, the visionary CEO and co-founder of Ryght, to talk about copilots and the application of generative AI in life sciences. Ryght is dedicated to revolutionizing the field of life sciences through the power of AI. By leveraging cutting-edge technology and innovative solutions, Ryght aims to empower professionals and organizations within the life sciences industry to streamline processes, enhance productivity, and drive meaningful outcomes.In our conversation, we discuss Simon's entrepreneurial background, the various companies he has founded, and what led him to create Ryght. We delve into the pivotal role of enterprise-scale, secure AI solutions in healthcare, and learn how Ryght's platform is reshaping the landscape of drug development and clinical research. Discover the intricate workings of generative AI copilots, the challenges of minimizing hallucinations and validating AI models, and why the utility of the approach at the enterprise level is essential. Simon also shares Ryght’s long-term goals and invaluable advice for leaders of AI startups. Join us, as we explore a world where healthcare and life sciences are transformed by cutting-edge technology with Simon Arkell from Ryght!Key Points:Hear about Simon’s background and his path to founding Ryght.Ryght’s generative AI approach, its potential in life sciences, and the role of copilots.The importance of enterprise-scale, secure AI solutions in healthcare.How generative AI copilots accelerate drug development processes.Differences between training models for life sciences versus generic AI models.Discover the challenges encountered in AI-powered solutions.Explore the company’s approach to customer feedback and model validation.Strategic considerations and advice for leaders of AI startups.Ryght’s mission to transform the healthcare and life sciences industry.Where to find more information about Ryght and connect with Simon.Quotes:“We built an enterprise-secure version of Generative AI that has many different features that allow large companies and small companies to very securely benefit from Generative AI without all of the issues that a very insecure, non-industry-trained solution might create.” — Simon Arkell“With this type of [generative AI] technology, you have the ability to completely unlock new formulas, and new molecules that could be life-changing.” — Simon Arkell“Improving the utility of the platform comes down to the efficacy of the output. It comes down to the in-context learning, the ensembling, and the prompting. But at the end of the day, a human has to determine, in many cases, the accuracy and relevance of a specific answer.” — Simon Arkell“It's not really about building models. It's about making sure that the right models are being utilized for the copilot.” — Simon ArkellLinks:Simon Arkell on LinkedInRyghtRyght on LinkedInRyght on XRyght on YouTubeResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

AI in healthcare is one of the most researched areas today, particularly on the clinical side of healthcare. Sean Cassidy is the Co-Founder and CEO of Lucem Health. Having worked in digital health for the last twenty years, he joins me today to talk about identifying chronic diseases. Tune in to hear how AI and machine learning are creating efficiencies for different forms of healthcare data, and how changes and challenges are being addressed to improve the process. Going beyond workflow support, we discuss considerations to bear in mind when integrating AI into healthcare systems and how to meaningfully measure efficacy in a clinical context. Sean shares some hard-earned wisdom about leading an AI startup, reveals his big vision for the future of Lucem Health, and much more.Key Points:Introducing guest Sean Cassidy, who co-founded Lucem Health. Defining digital health through an overview of Sean’s history in this industry. The founding idea behind Lucem Health. Different forms of healthcare data and how AI and machine learning can support them. Navigating changes in external variables and patient circumstances.The downstream diagnosis process and why patients are rarely re-assessed.How Lucem Health’s approach facilitates doctors as they continue as they always have. Considerations to bear in mind with the clinical adoption of AI beyond workflow.How efficacy is measured in a clinical context.Advice for leaders in AI startups.A vision for the future of Lucem Health. Quotes:“We are focused on early disease detection almost exclusively, and so that is using AI and machine learning algorithms to, at any point in time, evaluate the risk that a patient may have a certain disease.” — Sean Cassidy“Workflow is really important, but there are also other considerations that matter in terms of AI being more widely adopted in clinical settings and healthcare.” — Sean Cassidy“We are always evaluating and trying to get a deep understanding of whether what we said was going to happen with respect to the performance of the solution is actually manifesting itself in the real world.” — Sean CassidyLinks:Sean Cassidy on LinkedInLucem HealthResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

In this episode, I sit down with Jean-Baptiste Schiratti, Medical Imaging Group Lead and Lead Research Scientist at Owkin, to discuss the application of self-supervised learning in drug development and diagnostics. Owkin is a groundbreaking AI biotechnology company revolutionizing the field of medical research and treatment. It aims to bridge the gap between complex biological understanding and the development of innovative treatments. In our conversation, we discuss his background, Owkin's mission, and the importance of AI in healthcare. We delve into self-supervised learning, its benefits, and its application in pathology. Gain insights into the significance of data diversity and computational resources in training self-supervised models and the development of multimodal foundation models. He also shares the impact Owkin aims to achieve in the coming years and the next hurdle for self-supervised learning.Key Points:Introducing Jean-Baptiste Schiratti, his background, and path to Owkin.Details about Owkin, its mission, and why its work is significant.The application of self-supervised learning in drug development and diagnostics.Examples of the different applications of self-supervised learning.Discover the process behind training self-supervised models for pathology.Explore the various benefits of using self-supervised learning.His approach for structuring the data used for self-supervised learning.Unpack the potential impact of self-supervised AI models on pathology.Gain insights into the next frontier of foundation model development.He shares his hopes for the future impact of Owkin.Quotes:“To be able to train efficiently, computer vision backbones, you actually need to have a lot of compute and that can be very costly.” — Jean-Baptiste Schiratti“There are some models that are indeed particular to specific types of tissue or specific sub-types of cancers and also the models can have different architectures and different sizes, they come in different flavors.” — Jean-Baptiste Schiratti“The more diverse the [training] data is, the better.” — Jean-Baptiste Schiratti“I’m convinced that the foundation models will play a very important role in digital pathology and I think this is already happening.” — Jean-Baptiste SchirattiLinks:Jean-Baptiste Schiratti on LinkedInJean-Baptiste Schiratti on XOwkinPhikonResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Joining me today are Linda Chung and Michael Mullarkey to discuss the transformative potential of AI in mental health care. Linda is the co-CEO and Co-Founder of Aiberry, a groundbreaking AI company redefining mental healthcare accessibility. With a background in speech-language pathology, Linda pioneered telehealth services and now leads Aiberry in leveraging innovative technology for objective mental health screenings. Michael, the Senior Clinical Data Scientist at Aiberry, is dedicated to translating complex data science into tangible human value. His unique background in clinical psychology merged with a passion for coding drives his mission to address pressing human concerns through data.In our conversation, we explore the fascinating intersection of clinical expertise and artificial intelligence, unlocking personalized insights and proactive strategies for mental well-being. Hear about Aiberry’s innovative chatbot “Botberry” and how it helps provide insights into the user’s mental health. We also get into the weeds and unpack how Aiberry develops its models, data source challenges, the value of custom models, mitigating model biases, and much more! Our guests also provide invaluable advice for other startups and share their future vision for the company. Tune in and discover AI technology at the forefront of mental health innovation with Linda Chung and Michael Mullarkey from Aiberry!Key Points:Linda’s healthcare background and motivation for starting Aiberry.Michael's transition from clinical psychology to AI at Aiberry.Aiberry's AI-powered mental health assessment platform and its unique approach.The role of machine learning in Aiberry's technology.Model development, data collection challenges, and custom model creation.Addressing bias in models trained on patient interview data.Measuring impact and success metrics at AiberryAdvice for leaders of AI-powered startups.The vision for Aiberry's impact in the next three to five years.Quotes:“We know that early detection leads to early intervention and better outcomes.” — Linda Chung “Our models take the messy, natural human way that people talk about their mental health, and we turn it into systematic data that are necessary for the healthcare industry and report it back to the user.” — Linda Chung“As a health tech company, we have to take the health and the tech elements of our business equally seriously. So, one of our guiding principles from a health perspective is we have to keep people's data secure.” — Michael Mullarkey“We really value letting people talk about their mental health in their own words, and that can lead to some unexpected outcomes on the modeling side of the operation.” — Michael MullarkeyLinks:Linda Chung on LinkedInMichael Mullarkey on LinkedInAiberryResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Machine learning can be used as an innovative method to contribute to climate change resiliency. Today on Impact AI, I am joined by the co-founder and CTO of Vibrant Planet, Guy Bayes, to discuss how they are using AI to revitalize forests. Listening in, you’ll hear all about our guest’s background, why he started Vibrant Planet, what the company does, how they apply machine learning to their work, and a breakdown of how they collect the four sets of data they need. We delve into any problem areas they face in their individual and integrated data types before Guy tells us how they cross-validate their models. We even talk about how the teams collaborate, how machine learning and forest knowledge come together, and where he sees the company in the next three to five years. Finally, our guest shares some pearls of wisdom for any leaders of AI-powered startups.Key Points:A warm introduction to today’s guest, Guy Bayes. Guy tells us about his background and what led him to create Vibrant Planet. What Vibrant Planet does and how it contributes to climate change resiliency. How Vibrant Planet applies machine learning to the work.A breakdown of the four sets of data they need and how they collect it. The challenges they face when it comes to collecting and integrating all their data. How Guy makes sure that their models work in different geographic regions. Incorporating forest knowledge into data modeling and machine learning development. How the Vibrant Planet teams work together and collaborate to achieve their goal. What Vibrant Planet does to measure the impact of this technology. New AI advancements Guy is particularly excited about for Vibrant Planet. Guy shares some advice for leaders of AI-powered startups.Where he sees Vibrant Planet’s impact in the next three to five years. Quotes:“Getting the forest back into a state that's more able to tolerate fire and more able to produce low-intensity fire rather than high-intensity fire is [Vibrant Planet’s] goal.” — Guy Bayes“We have – not only super good engineers but also very talented ecological scientists and people that have done physical hands-on forestry for their careers. – This mix of those three personas – work together pretty harmoniously actually because we all share a common goal.” — Guy Bayes“I don't think you can ever find one person who has all that in their head, but you can find a team that does.” — Guy Bayes“You will not have an impact without having a combined team that all respects each other and brings different things to the table.” — Guy BayesLinks:Guy Bayes on LinkedInVibrant PlanetVibrant Planet on LinkedInResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

In this episode, I sit down with Erez Naaman, co-founder and CTO of Scopio Labs, to delve into the transformative potential of AI in healthcare, particularly in blood cell morphology analysis. Erez shares the intriguing journey behind the inception of Scopio Labs which was driven by a desire to revolutionize healthcare practices. Discover how Scopio Labs' platforms digitize and streamline the process of blood cell analysis and the pivotal role of machine learning in distinguishing and classifying various cell types. Gain insights into the significance of data collection and algorithm development, the evolution of AI infrastructure over the past decade, regulatory considerations on product development, and more. He also shares invaluable insights for AI startup leaders, the future trajectory of Scopio Labs, and the profound impact envisioned for the healthcare landscape. Join me as we explore the intersection of AI and healthcare innovation with Erez Naaman.Key Points:Eres shares his professional background and his path to founding Scopio Labs.Revolutionizing healthcare through AI-driven blood cell morphology analysis.The pivotal role machine learning plays in distinguishing and classifying various cell types.Discover the challenges of working with blood smear images; particularly for training models.Learn about the differences between regulated and nonregulated machine learning.AI infrastructure development and the associated regulatory considerations.Explore his approach to developing new machine learning products or features. Hear why he chooses to prioritize the end-user experience during development.Advice for budding entrepreneurs and the future trajectory of Scopio Labs.Quotes:“In terms of the approach [to AI], I think we saw it the same way that we do today in terms of its importance but I think that the infrastructure for using ML has greatly evolved.” — Erez Naaman“Getting a large enough data set to get a reliable classification on specific more rare cell types is the most difficult problem in my opinion.” — Erez Naaman“In a way, we look at it backward. Machine learning is a tool and not a goal. So, we always start with the patient in mind or the user.” — Erez Naaman“Everyone is dealing with AI and so the front runners are clearly becoming the leaders with time. So, it is much easier to choose the right tools for every task as time progresses.” — Erez NaamanLinks:Erez Naaman on LinkedInScopio LabsResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Customer service calls often start and end at the operator’s headset, but there is so much untapped data from these conversations that could be used to improve business systems on a holistic level. Today’s guest, Amy Brown has seen the value of unlocking conversational data to improve healthcare systems across the country, and as the Founder and CEO of Authenticx, she has taken giant strides towards accomplishing this goal.Authenticx is an AI-powered platform that makes it possible for healthcare organizations to have a single source of conversational data, creating powerful and immersive customer insight analysis that informs business decisions. In today’s conversation, Amy explains why she founded Authenticx, what the company does, and why her business is important for healthcare.  We also learn about how the company uses machine learning in its processes, the challenges of working with conversational data, how Authenticx upholds a high ethical standard, and how the impact of its technology can be measured across healthcare systems nationwide. After sharing some important advice for other leaders of AI-powered startups, Amy explains why Authenticx will be a key player in healthcare for the foreseeable future. Key Points:A warm welcome to the Founder and CEO of Authenticx, Amy Brown. Amy’s professional background, and how she ended up founding Authenticx. What Authenticx does and why the company is important for healthcare. How the company uses machine learning to get better insights from conversational data.  A closer look at the conversational data that Authenticx works with. The challenges of working with and training models on conversational data.  Other ways that they validate their models. Mitigating biases and upholding ethics. How Amy measures the impact of Authenticx’s technology.Her advice to other leaders of AI-powered startups. Where Authenticx will be in the next three to five years, according to Amy.Quotes:“That’s really what I’m trying to get at; using technology to help explain customer and consumer perception of their care, and using that; putting that to work for the healthcare industry so it can start to improve its systems in a way that allows patients and consumers to actually get a better outcome.” — Amy Brown“Our data team has had to become extremely proficient at dealing with all kinds of messy data.” — Amy Brown“We’ve hired a diverse group of human beings because we want to make sure that we’re inclusive in our interpretations of what’s happening in these conversations.” — Amy Brown“You can never eliminate all bias – we would never purport of doing that – but we can be very intentional about how we train the data.” — Amy Brown“[The] dream scenario is that the healthcare system in this country starts to make room for and evolve in how it makes its business decisions to include the voices of their customers as a key source of insight, intel, and data.” — Amy BrownLinks:Amy Brown on LinkedInAmy Brown on XAuthenticxAuthenticx on InstagramResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Using biological intelligence, human intelligence, and artificial intelligence, the company in the spotlight today aims to demystify health, make science accessible, and honor the biochemical individuality of every human.Today on Impact AI, I am joined by the founding CTO and Head of Discovery AI at Viome, Guru Banavar! He is here to talk all about AI and the human microbiome. As you tune in, you’ll hear about Guru’s background and what led to the creation of Viome, including what they do and why their work is crucial to chronic disease. He unpacks their use of machine learning to turn RNA data into insights for their customers, the challenges they face in training models for the work they do, and Guru sheds light on the early steps of their process for planning and developing new machine-learning products or features.  Be sure not to miss out on this insightful conversation about how Guru and the team at Viome are working to decode the human microbiome.Key Points:Learn about Guru’s background and what led to the creation of Viome.What Viome does and why it’s important for chronic disease.Using machine learning to turn RNA data into insights for customers at Viome.Guru highlights the challenges they face in training models based off of the work with RNA data and the large data set they’ve collected from customers. He unpacks the early steps in the process of planning and developing a new machine-learning product or feature.We talk about technological advancements that made it possible to build their technology. Guru’s advice to other leaders of AI-powered startups.His thoughts on the impact of Viome in the next 3-5 years.Quotes:“At some point in time, I decided that the impact that I wanted to make in the field of computational biology, life sciences, and healthcare could be done only if I joined a few of my friends from the broader community, and started a new company — [Viome].” — Guru Banavar“I am one of those AI people who believes that you first focus on the problem, and you bring all of the tools you need to solve the problem. AI, to me, is not just one thing, like the latest buzzword. For me, AI is an ML, a set of tools, and you take the right tool for the right problem.” — Guru Banavar“One of our core intellectual property elements is the meta-transcriptomic laboratory technology, which essentially, isolates, detects, and processes what we call the informative RNA molecules in any given sample. That required a number of sort of biochemistry-level technology breakthroughs.” — Guru Banavar“I would advise other leaders of AI-powered startups to be very careful about how you pick your solution toolset, based upon the problem that you want to solve.” — Guru BanavarLinks:Guruduth Banavar on LinkedInGuruduth Banavar on XViomeViome BlogResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

In this episode, I sit down with Konstantinos Kyriakopoulos, CEO of DeepSea, to discuss the transformative world of AI-powered shipping optimization. DeepSea focuses on enhancing vessel performance, fuel efficiency, and overall logistics management in the shipping and logistics industry. Konstantinos has been a key figure in advocating for digitalization in the maritime sector, pushing for technologies to streamline processes, cut costs, and reduce environmental impact.In our conversation, Konstantinos shares the captivating journey behind DeepSea's inception, revealing how its AI-driven solutions emerged from a desire to revolutionize the shipping industry's efficiency and environmental impact. We explore the intricate use of machine learning to predict fuel consumption, optimize vessel operations, and navigate the shift toward decarbonization.Gain insights into the intricacies of data architecture, the critical role of scalability, measuring impact, the future vision of the company, and much more. Don't miss out on discovering the cutting-edge applications of AI that are steering the shipping industry toward a more sustainable future with Konstantinos Kyriakopoulos. Tune in now!Key Points:Background about Konstantinos and DeepSea's inception.How AI is reshaping shipping efficiency and vessel operations.The role of DeepSea in the shipping industry and mitigating climate change.Insights into the challenges and hurdles of an evolving shipping industry.How DeepSea leverages AI, inputs into the model, and the overall aim.Approaches the company implements to ensure the integrity of its products.Why the explainability of machine learning models is critical. He shares DeepSea’s approach to model validation.Measuring impact: CO2 reduction and cost savings for clients.Konstantinos offers valuable advice for leaders of AI-powered startups.What the company has planned for the future.Quotes:“If you really want to create impact, it’s not enough to just show people what’s happening and give them analytics, but you also have to, in some way, produce a tangible ROI.” — Konstantinos Kyriakopoulos“The most important thing is to evaluate performance, so to make sure that the proof of performance is constantly being tested and you have good benchmarks and analytics.” — Konstantinos Kyriakopoulos“It’s really important to also be able to check internally what is going on but also how the customer wants to see what’s created.” — Konstantinos Kyriakopoulos“For us, the impact is actually very straightforward. It’s dollars and the metrics tonnes of CO2.” — Konstantinos Kyriakopoulos“I think what I always say when people talk to me about starting an AI company is to focus on your data architecture early.” — Konstantinos KyriakopoulosLinks:Konstantinos KyriakopoulosDeepSeaDeepSea on LinkedInResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Foundation Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

AI and machine learning have had a huge impact on the healthcare industry, but there are still plenty of advances to be made. Joining me today is Sam Rusk, Co-founder and CAIO of EnsoData, to talk about how their team is using machine learning to optimize sleep. Tuning in, you’ll learn about the founding of EnsoData, their implementation of ML, and the important role they play in the healthcare sector. We discuss the primary challenges of working with and training models on waveform data, EnsoData’s diagnostic processes, and how they use ML to process collected waveforms and identify therapy opportunities. Sam also shares his thoughts on how ML has developed since they first founded the company nine years ago, his advice for other leaders of AI-powered startups, and what his hopes are for EnsoData in the next five years. To learn how EnsoData is making waves in healthcare, be sure to listen in today!Key Points:Sam’s engineering and entrepreneurship background and EnsoData’s origin story.What EnsoData does and why it’s important for healthcare.Using ML to process collected waveforms and identify therapy opportunities.Input and output models EnsoData uses to navigate the noise of tricky signal types.Examples of what they are trying to predict with these models.Diagnostic processes used in sleep medicine and the role of EnsoData.Major challenges of working with and training models on waveform data.Different approaches EnsoData has implemented to tackle generalizability.Ways that the role of ML has evolved since EnsoData was founded nine years ago.Insight into their team’s process for developing new products and features.EnsoData’s place in the clinical workflow and how they assist doctors and patients.Sam’s advice for other leaders of AI-powered startups.What’s next for EnsoData and where you can go to learn more!Quotes:“We have a pretty mature process for taking feature ideas and moving them from the top of the funnel on product management all the way to testing and releasing those.” — Sam Rusk“We spend a lot of our time solving not necessarily the machine learning performance side of the problem, but more ‘how do we get this into the clinicians’ hands in a way that makes sense for everyone.’” — Sam Rusk“While we want to deliver products that change the game, we [also] invest heavily in research, and we are active in the community, publishing and engaging in the research community in sleep.” — Sam RuskLinks:Sam Rusk on LinkedInEnsoDataResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Custom Vision Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

What if you were told that AI could improve agriculture, reduce climate change, and potentially solve global food insecurity? In this episode of Impact AI, I am joined by Ranveer Chandra from Microsoft Research to discuss his work in the world of agriculture. Tuning in, you’ll hear all about Ranveer’s career, how he got his agriculture idea picked up by Microsoft, data-driven agriculture, and more! We then delve into the data needed to achieve their goals before Ranveer discusses all the challenges they face when it comes to multimodal AI. Ranveer is very hopeful that machine learning can drastically improve agriculture. He tells me what new AI technologies he is most excited about, their potential impact on agriculture, and even shares advice for other leaders in AI. Finally, my guest warns us against the potential divide society can create if AI is not made accessible to all people. You don’t want to miss out on this informative and incredibly interesting episode so press play now!Key Points:Introducing today’s guest, Ranveer Chandra.A bit about Ranveer’s background and how he landed up at Microsoft Research. How Microsoft got involved in agriculture. Ranveer tells us about data-driven agriculture, what it means, and how he plans to achieve it. The kinds of data they collect from farms in order to achieve these goals. Challenges associated with multimodal AI.How these technologies have been deployed so far.  What new technology Ranveer is excited about in the world of machine learning.Ranveer shares some advice for other leaders of AI-based products. The potential impact of data-driven and AI technologies for agriculture in the future. Ranveer warns us about the dangers of creating an AI-divide and what that would mean. Quotes:“Technology could have a deep impact on agriculture. It could address the world's food problem; it could help improve livelihoods of a lot of smallholder farmers.” — Ranveer Chandra“The key question is, how do you sustainably nourish the planet? How do you sustainably nourish the people in this world?” — Ranveer Chandra“Microsoft is not an agriculture company. So we are not sending anything to farmers, but we are providing the tools on top of which you could build solutions for farmers, or partners, or customers build solutions and take the solutions to farmers.” — Ranveer Chandra“We need to make data consumable, and generative AI has the suitability to make that data more consumable.” — Ranveer Chandra“There are over 500 million smallholder farmers worldwide whose lives would benefit with artificial intelligence.” — Ranveer ChandraLinks:Ranveer Chandra on LinkedInRanveer Chandra on XRanveer Chandra on InstagramMicrosoft Research – Ranveer ChandraResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Custom Vision Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Continuous glucose monitors (CGMs) are a trusted tool for diabetics, but today’s guest believes that widespread adoption could also be valuable for reversing the obesity crisis. Meet Bill Tancer, the Co-founder and Chief Data Scientist of Signos, a metabolic health platform that combines CGMs with a unique AI engine to offer real-time data and recommendations for healthy weight management.Today, Bill joins me to talk about all things metabolic health and machine learning. Tune in as we discuss how the Signos team trains their machine learning algorithms, the challenges they encounter when it comes to gathering data, and some of the other external factors that influence the performance of their model. We also touch on the value of qualitative data in the form of user feedback, the importance of keeping your mission in mind in the rapidly expanding AI space, and so much more! To find out how Signos is unlocking metabolic health with ML, don’t miss this episode of Impact AI.Key Points:Reflecting on the personal and professional paths that led Bill to create Signos.What Signos does for glycemic dysregulation and why it’s so important for healthcare.Insight into the role that ML plays in Signos’ technology.How Signos trains their ML algorithms using various sources of data.Food logging and other challenges that come with gathering CGM data.Ways that external factors influence model performance and how Signos mitigates that.Qualitative user responses that help Bill measure the impact of this technology.Bill’s mission-driven advice for other leaders of AI-powered startups.How he believes the impact of Signos will continue to evolve going forward.Quotes:“Along with diabetes as its own health risk, having [dysregulated] glucose can lead to other medical problems. Cardiovascular disease, stroke, Alzheimer's, just to name a few. [It] is such an important goal for [Signos] to help people reduce their glycemic variability.” — Bill Tancer“That's what gets me up in the morning; hearing [positive user anecdotes]. That, in conjunction with looking at our own data and how our members are improving in terms of their wellness, tells us we're having a measurable impact.” — Bill Tancer“It is so easy [with] all the things you can do with AI to end up in a space where you've got a solution that's searching for a problem to solve. The antidote to finding yourself in that situation is always returning back to your mission.” — Bill TancerLinks:SignosBody Signals PodcastBill Tancer on LinkedInBill Tancer on InstagramResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Custom Vision Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

In an emergency setting, making a quick diagnosis under pressure is often a matter of life or death. This is especially true when it comes to diagnosing infectious diseases. Unfortunately, diagnosing infections in an emergency department is rife with challenges. Current tests either take too long, deliver unreliable results, or both. That’s where Inflammatix comes in. They are using machine learning technology to develop a point-of-care instrument that will diagnose the type of infection, and severity of infection, in emergency care quickly and effectively. Their first main product is currently in the late stages of development and can deliver a test report in about half an hour using cold blood as a sample source.Joining me today to shed light on this incredible initiative is Ljubomir Buturovic, Vice President of Machine Learning at Inflammatix. We hear from Ljubomir about the role that machine learning played in this technology, key challenges they’ve encountered while training models on gene expression data, how they selected the 29 clinically relevant genes based on published scientific papers, plus a whole lot more. Tune in today to learn more about the groundbreaking work being done at Inflammatix and what you can expect from them in future!Key Points:A warm welcome to today’s guest Ljubomir Buturovic.Ljubomir’s background in machine learning and what led him to Inflammatix.An overview of the important work being done at Inflammatix in healthcare.Details about their main product for diagnosis in emergency care.The role of machine learning in their technology to measure gene expression.How they selected the 29 clinically relevant genes based on published scientific papers.Key challenges they encountered while training models on gene expression data.Ground truth labels; the strategies they used to identify infections and validate their models.How they made sure that their models would work for multiple assay platforms.Using grouped analysis to ensure their models would serve a diverse patient population.Their approach to developing technology that would fit in with the clinical workflow and provide the right assistance to doctors and patients.The benefits that Inflammatix has seen from publishing their work.Ljubomir’s advice to other leaders of AI-powered startups working in healthcare.Where you can expect to see Inflammatix in five years.Quotes:“We developed an instrument which measures this gene expression for 29 clinically relevant genes for infections.” — Ljubomir Buturovic“It takes a long time to achieve adoption. This is basically applying AI in medicine. When you are applying AI in medicine, the whole process of development and adoption works on medicine timescales, not on AI timescales.” — Ljubomir Buturovic“One of the key challenges in applying machine learning in clinical test design is the availability of samples for training and validation. This is in sharp contrast to other applications, like maybe movie recommendations, or shopping, where you have a lot of input data, because it's relatively easy to collect.” — Ljubomir ButurovicLinks:InflammatixInflammatix's Machine Learning BlogLjubomir Buturovic on LinkedInLjubomir Buturovic on XResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Computer Vision Advisory Services – Monthly advisory services to help you strategically plan your CV/ML capabilities, reduce the trial-and-error of model development, and get to market faster.

There are now a few different AI foundation models available for Earth Observation (EO) data. These vast neural networks can be rapidly fine-tuned for many downstream tasks, making them a highly versatile and appealing tool.Today on Impact AI, I am joined by Hamed Alemohammad, Associate Professor in the Department of Geography at Clark University, Director of the Clark Center for Geospatial Analytics, and former Chief Data Scientist of the Radiant Earth Foundation, to discuss the applications of foundation models for remote sensing. Hamed’s research interests lie at the intersection of geographic information science and geography, using observations and analytical methods like machine learning to better understand the changing systems of our planet.In this episode, he shares his perspective on the myriad purposes that foundation models serve and offers insight into training and fine-tuning them for different downstream applications. We also discuss how to choose the right one for a given project, ethical considerations for using them responsibly, and more. For a glimpse at the future of foundation models for remote sensing, tune in today!Key Points:A look at Hamed’s professional journey and the research topics he focuses on today.Defining foundation models and the purposes they serve.The vast amount of data and resources required to train and fine-tune a foundation model.Ways to determine whether or not a foundation model will be beneficial.How foundation models improve generalizability for downstream tasks.Factors to consider when selecting a foundation model for a given downstream task.Insight into the future of foundation models for remote sensing.Hamed’s advice for machine learning teams looking to give foundation models a try.His take on the impact of foundation models in the next three to five years.Ethical considerations for the responsible use of AI that apply to foundation models too.Quotes:“[Foundation models] are pre-trained on a large amount of unlabeled data. Secondly, they use self-supervised learning techniques – The third property is that you can fine-tune this model with a very small set of labeled data for multiple downstream tasks.” — Hamed Alemohammad“It takes a lot to train a model, but you would not [do it] as frequently as you would [fine-tune] the model. You can use shared resources from different teams to do that - share it as an open-source model, and then anybody can fine-tune it for their downstream application.” — Hamed Alemohammad“The promising future [for foundation models] will be combining different modes of data as input.” — Hamed Alemohammad“There is a lot to do and the community is eager to learn, so if people are looking for challenging problems, I would encourage them to explore [the foundation model domain] and work with domain experts.” — Hamed AlemohammadLinks:Hamed AlemohammadHamed Alemohammad, Clark University Hamed Alemohammad on LinkedInHamed Alemohammad on XHamed Alemohammad on GitHubFoundation Models for Generalist Geospatial Artificial IntelligencePrithvi-100M on Hugging FaceHLS Multi-Temporal Crop Classification Model on Hugging FaceResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Custom Vision Model Assessment – Foundation models are popping up everywhere – do you need one for your proprietary image dataset? Get a clear perspective on whether you can benefit from a domain-specific foundation model.

Could there be a future where not using AI is considered unethical? With the growing efficiency created by AI support, radiologists are able to focus on the most important aspects of their work. During this conversation, I am joined by Stefan Bunk and Christian Leibig from Vara. Tuning in, you’ll hear about the essential practice of maintaining a high standard of data quality and how AI technology is revolutionizing breast cancer detection and treatment. We discuss the relevance of German innovation and research on a global community, and the step-by-step process that Vara adopts to test and introduce AI products. You’ll also hear about Stefan and Christian’s vision for the future of Vara. Don’t miss this episode, packed with powerful insights!Key Points:Introducing Stefan Bunk and Christian Leibig from Vara. Vara’s mission for breast cancer outcomes in line with WHO’s Global Breast Cancer Initiative.The role of machine learning in Vara’s technology.What the AI technology predicts and the software that goes into this. Why it is essential to maintain a high standard of data quality.The relationship between images from earlier exams and current procedures. How models are trained to manage different variations. The relevance of German data for global application.Why it is important to have strong processes around AI deployment. What it means to run in Shadow Mode first and why Vara chooses to do this with AI products.How they established the best way to integrate AI into the workflow.The crucial role of trust in machine learning models. Monitoring AI models constantly and creating the means to react quickly.Where Stefan and Christian see the impact of Vara in five years. The enduring goal of Vara: to support radiologists as they focus on the most important factors. Considering the possibility that not using AI will become unethical in the future. Quotes:“Our ambition is to find every deadly breast cancer early. Breast cancer is the most common cancer actually worldwide, one out of eight women will have it at some point in their lifetime.” —  Stefan Bunk“At Vara, we want to empower health systems to systematically find more cancers much earlier and systematically downstage cancers.” — Stefan Bunk“A machine learning model can actually outperform a radiologist with a single image, but nevertheless, can still benefit from taking comparisons across images into account.” —  Christian Leibig“When you roll out a technology such as AI, which is the technology that is hard to understand, and you cannot always predict how it behaves in certain edge cases. We believe there must be strong processes around it wherever you will deploy your AI.” — Stefan BunkLinks:Stefan Bunk on LinkedInChristian Leibig on LinkedIn VaraResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Computer Vision Advisory Services – Monthly advisory services to help you strategically plan your CV/ML capabilities, reduce the trial-and-error of model development, and get to market faster.

Ready to dive deep into the world of online security and identity verification? In this episode, I sit down with Vyacheslav Zholudev from Sumsub to discuss user verification, fraud detection, and the role of machine learning in ensuring the safety of digital interactions. Vyacheslav is the co-founder and CTO of Sumsub, an online verification platform that secures the whole user journey using innovative transaction monitoring and fraud prevention solutions.In our conversation, Vyacheslav discusses the evolution of Sumsub, its role in online identity verification, and the challenges posed by deepfakes in the digital world. We explore the cat-and-mouse game against the rising threat of deepfakes, the pivotal role of machine learning in user verification, the challenges posed by generative AI advancements, the ethical considerations in combating biases, and much more. Tune in and discover the future of user verification with Vyacheslav Zholudev from Sumsub!Key Points:Vyacheslav's background and the journey that led to the creation of Sumsub.Evolution of Sumsub from an anti-Photoshop project to a user verification platform.Hear why online user verification is vital for implementing digital features.Sumsub’s overall mission and shifting from physical to online identity verification.The crucial role of machine learning in Sumsub’s user verification technology.How the latest generative AI advancements impact user verification efficiency.Implications of deepfakes on society and their potential to facilitate fraud.Approaches and techniques used by Sumsub to detect and combat deepfakes.Continuous learning and adaptation in the rapidly evolving field of machine learning.Ethical concerns and potential biases in models trained for fraud detection.Monitoring and preparing to address potential bias in Sumsub’s models.Advice for leaders of AI-powered startups and Sumsub's future goals.Quotes:“Basically, [machine learning] is everywhere. I can’t imagine that our company could exist without machine learning and different algorithms in this area.” — Vyacheslav Zholudev“It was really expensive and difficult to create a deepfake that looks realistic. Nowadays, you can do it with a click of a button on your smartphone. That became a problem [for user verification].” — Vyacheslav Zholudev“We have a very strong machine learning team and we’re really focusing a lot nowadays on fighting those deepfakes, trying new and new ways how we can protect ourselves and our customers against them.” — Vyacheslav Zholudev“Think like a hacker and don’t compromise security. Don’t think that some things won’t be revealed, they will.” — Vyacheslav ZholudevLinks:Vyacheslav Zholudev on LinkedInSumsubResources for Computer Vision Teams:LinkedIn – Connect with Heather.Computer Vision Insights Newsletter – A biweekly newsletter to help bring the latest machine learning and computer vision research to applications in people and planetary health.Computer Vision Strategy Session – Not sure how to advance your computer vision project? Get unstuck with a clear set of next steps. Schedule a 1 hour strategy session now to advance your project.Computer Vision Advisory Services – Monthly advisory services to help you strategically plan your CV/ML capabilities, reduce the trial-and-error of model development, and get to market faster.