Super Data Science: ML & AI Podcast with Jon Krohn

Host Jon Krohn unpacks Dario Amodei’s vision of a techno-utopia in his essay Machines of Loving Grace, where “Powerful AI” takes center stage. Amodei, CEO of Anthropic, imagines a future where AI doesn’t just assist but actively shapes fields like healthcare, economics, and governance with unmatched intelligence and autonomy. Jon explores the possibilities and challenges of this AI-driven future, asking how close we are to seeing these revolutionary shifts and what they mean for society. Additional materials: www.superdatascience.com/832 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

PyTorch Lightning is revolutionizing the AI landscape, and Dr. Luca Antiga, CTO of Lightning AI, joins host Jon Krohn to explain how. In this episode, they explore the tools pushing AI development forward, from Lightning Studios to Lit-Serve, and discuss the game-changing rise of small language models that challenge industry giants with precision and speed. Luca also shares his vision for developers in an AI-enhanced world, where coding meets creativity and collaboration with intelligent tools. This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick, and by ODSC, the Open Data Science Conference. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: How Lightning AI's open-source tools make AI development faster [11:30] The rise of small language models and how they'll rival LLMs [37:47] Luca's journey from biomedical imaging to deep learning pioneer [52:03] How AI will transform software developer tasks [1:03:05] Additional materials: www.superdatascience.com/831

Geoffrey Hinton and Sir Demis Hassabis: The Nobel Prize committee is an achievement of the highest order, awarding physicists, chemists, physiologists, medical practitioners, writers, pacifists and economists perhaps the greatest honor in their respective fields. In this week’s Five-Minute Friday, Jon Krohn discusses how two AI pioneers came to win prizes in chemistry and physics. Additional materials: www.superdatascience.com/830 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Neuroscientist Bradley Voytek outlines to Jon Krohn the incredible use of data science and machine learning in his research and how recent discoveries in action potentials and neurons have completely skyrocketed the field to a new understanding of the brain and its functions. You’ll also hear what Bradley thinks is most important when hiring data scientists and his contributions to Uber’s algorithm when it was still a startup.  This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick, and by Gurobi, the Decision Intelligence Leader. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: Breakthroughs in brain region communication [04:08] The future of brain research and MedTech [35:24] The libraries and software used at the Halicioglu Data Science Institute [45:11] Brain rhythm as a diagnostic tool [1:02:58] Bradley’s curriculum structure at UC San Diego [1:12:21] How Uber applies data science [1:20:07] Additional materials: www.superdatascience.com/829

The citizen data scientist: Fact or fiction? Jon Krohn holds a conversation across episodes in this Five-Minute Friday, with today’s guest Keith McCormick, in part responding to Nick Elprin’s interview in episode 811: Scaling Data Teams Effectively. Additional materials: www.superdatascience.com/828 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Ritchie Vink, CEO and Co-Founder of Polars, Inc., speaks to Jon Krohn about the new achievements of Polars, an open-source library for data manipulation. This is the episode for any data scientist on the fence about using Polars, as it explains how Polars managed to make such improvements, the APIs and integration libraries that make it so versatile, and what’s next for this efficient library. This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick, by Gurobi, the Decision Intelligence Leader, and by ODSC, the Open Data Science Conference. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: Why Polars is so efficient [05:20] Polars’ easy integration with other data-processing tools [21:23] Eager vs lazy executive in Polars [32:15] Polars’ data processing of large- and small-scale datasets [38:28] Ritchie’s plans to scale his company [46:14] Upcoming features in Polars [58:06] Additional materials: www.superdatascience.com/827

Next-gen IDEs, efficiency-boosting open-source Python libraries, and changes in hiring for data scientists: This episode of In Case You Missed It gives you our best clips of September’s interviews, hosted by Jon Krohn. Additional materials: www.superdatascience.com/826 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Data contracts are redefining data quality and governance, and Chad Sanderson, CEO of Gable.ai, joins host Jon Krohn to explain how they can transform your data strategy. He breaks down what data contracts are, how they shift data quality checks closer to production, and why they’re essential for reducing data debt. Chad also highlights how better alignment between data producers and consumers can elevate data reliability and tackle change-management challenges in modern organizations. This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick, and by Gurobi, the Decision Intelligence Leader. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: What data contracts are and how they define expectations for data quality [03:16] What data contracts look like [09:09] The common misconceptions about data quality when implementing AI [12:55] Chad’s Chief Operator role at Data Quality Camp [19:46] How “shifting left” improves data reliability by addressing issues early [24:17] Why data professionals still struggle with data quality [30:31] How data debt forms and why it leads to complex, inefficient architectures [35:53] How will the role of human oversight evolve in ensuring data quality? [47:12] How can data teams leverage storytelling? [52:33] Additional materials: www.superdatascience.com/825

Llama 3.2 brings a new era of AI innovation with lightweight models tailored for on-device applications and powerful vision models for handling complex image inputs. Host Jon Krohn explores how this release pushes the boundaries of open-source AI, making it more accessible and versatile for developers. He also covers the Llama Stack toolkit, designed to streamline deployment, and Llama Guard 3, Meta’s latest content moderation solution. With extensive support from major cloud and hardware partners, Llama 3.2 is set to unlock groundbreaking possibilities for AI across mobile and beyond. Tune in to hear more. Additional materials: www.superdatascience.com/824 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Virtual humans are rewriting the rules of digital communication and reshaping entire industries. This week, Jon Krohn welcomes Natalie Monbiot, Head of Strategy at Hour One, to shed light on how AI avatars are revolutionizing L&D and e-commerce by turning traditional training and product listings into captivating, presenter-led content. This episode is brought to you by epic LinkedIn Learning instructor Keith McCormick, by Gurobi, the Decision Intelligence Leader, and by ODSC, the Open Data Science Conference. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • How do you create a virtual being? [10:55] • Reid Hoffman's avatar [13:40] • The virtual human economy [31:07] • Virtual human societies [51:24] • Virtual humans and creative expression [56:35] • Challenges in maintaining transparency [01:00:22] Additional materials: www.superdatascience.com/823

NotebookLM, Google’s latest AI tool, takes content creation to a new level. This week, Jon Krohn shares how the platform transformed his 200-page dissertation into a fascinating 11-minute podcast. Discover how AI can turn vast amounts of information into engaging and digestible content, opening up new possibilities for content creation. Additional materials: www.superdatascience.com/822  Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Marck Vaisman speaks to Jon Krohn about his paradigm for understanding core data practitioner types. Hear Marck detail the four data practitioner personas that he has identified in his research, why he believes the roadmaps that influencers like to promote as surefire ways to a data science career don’t work in practice, and why the term “data scientist” is still so elusive and hard to recruit for. This episode is brought to you by Gurobi, the Decision Intelligence Leader. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • How Marck started his work in defining data science roles [08:06] • The relationship between the four data practitioner personas [15:26] • About Marck’s “menu” for effective data science [40:43] • How recruiters can hire the best data scientist for the job [59:31] Additional materials: www.superdatascience.com/821

Jon Krohn takes OpenAI’s new models (o1-preview and o1-mini) for a spin in this Five-Minute Friday, learning their key strengths and limitations, and how the o1 series may represent yet another landmark for generative AI. Additional materials: www.superdatascience.com/820  Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

SuperDataScience veteran and Udemy teacher Luka Anicin is on the podcast to talk about his brand-new course, “PyTorch: From Zero to Hero”, available exclusively on superdatascience.com. Host Jon Krohn asks Luka why he feels that every data scientist should consider PyTorch as their default Python library, and why “keeping it simple” can secure the success of a machine learning project. This episode is brought to you by AWS Inferentia and AWS Trainium, and by Gurobi, the Decision Intelligence Leader. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • About the PyTorch library [03:29] • Why PyTorch became so popular [25:24] • How to increase accuracy and efficiency in PyTorch [31:49] • How to utilize transfer learning [35:44] • Why real-world projects are essential to data scientists [41:10] • About Datablooz [46:49] Additional materials: www.superdatascience.com/819

Experts from AI and data science discuss the impact and benefits of decentralization, the importance of structuring AI systems in business, and why knowing the basics will always matter for data engineers. Listen to Shingai Manjengwa (episode 809), Daniel Hulme (episode 807), Jerry Yurchisin (episode 813) and Nick Elprin (episode 811) explore a future world of work that rewards continuing learners, sets tasks for the people best suited to complete them rather than those whose job titles reflect the spec, and applies a fleet of ‘AI agents’ to solve complex business tasks. Additional materials: www.superdatascience.com/818  Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Dr. Julia Silge, Engineering Manager at Posit, introduces the brand-new Positron IDE, perfect for exploratory data analysis and visualization. She also lays out her top picks for LLMs that boost coding efficiency and discusses when traditional NLP methods might be the smarter choice over LLMs. Plus, Julia highlights some must-know open-source libraries that make managing MLOps easier than ever. Tune in for insights that every data scientist, ML engineer, and developer will find useful. This episode is brought to you by Gurobi, the Decision Intelligence Leader, and by ODSC, the Open Data Science Conference. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • Overview of Posit and Positron IDE [05:20] • How the needs of a data scientist differ from those of a software developer [10:54] • How to contribute to the open-source Positron [19:50] • MLOps and Vetiver: Tools for deploying and maintaining ML models [37:01] • Natural Language Processing (NLP) and the Tidyverse approach [50:34] • The role of AI and LLMs in data science education [1:24:18] Additional materials: www.superdatascience.com/817

Jon Krohn takes on a listener's challenge to explain his work in data science to his 94-year-old grandmother, Annie. This heartwarming conversation covers what data is, the role of a data scientist, and breaks down artificial intelligence (AI) and artificial general intelligence (AGI) in simple terms. The episode provides a fresh take on how to communicate complex topics to a lay audience, offering both clarity and insight. Additional materials: www.superdatascience.com/816  Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Polars, Python, Narwhals, Rust, and Pandas: Marco Gorelli talks to Jon Krohn about the many ways to use the newest data libraries available, the joys of open-source development, and the best method to win prizes in forecasting competitions. This episode is brought to you by AWS Inferentia and AWS Trainium, by Babbel, the science-backed language-learning platform, and by Gurobi, the Decision Intelligence Leader. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • When to use Polars vs Pandas [08:26] • How Polars optimizes string operations and data processing [20:08] • Where Narwhals outstrips Polars and Pandas [48:37] • The benefits of using Altair [55:21] • Addressing the lack of women in data science [1:09:58] • How to win a forecasting competition [1:16:58] Additional materials: www.superdatascience.com/815

As summer winds down, this episode shifts focus from the usual tech discussions to something more personal: reflecting on the importance of balancing work with life’s simple pleasures. While the world of data science and AI continues to evolve rapidly, it's essential to remember that true success isn't just about professional milestones. It’s also about cherishing the moments that make life meaningful. Tune in for a brief but impactful reflection on how to redefine success to include not just achievements, but also the everyday joys that often go unnoticed. Additional materials: www.superdatascience.com/814  Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.

Jerry Yurchisin from Gurobi joins Jon Krohn to break down mathematical optimization, showing why it often outshines machine learning for real-world challenges. Find out how innovations like NVIDIA’s latest CPUs are speeding up solutions to problems like the Traveling Salesman in seconds. Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: • The Burrito Optimization Game and mathematical optimization use cases [03:36] • Key differences between machine learning and mathematical optimization [05:45] • How mathematical optimization is ideal for real-world constraints [13:50] • Gurobi’s APIs and the ease of integrating them [21:33] • How LLMs like GPT-4 can help with optimization problems [39:39] • Why integer variables are so complex to model [01:02:37] • NP-hard problems [01:11:01] • The history of optimization and its early applications [01:26:23] Additional materials: www.superdatascience.com/813