Context Engineering in Practice: How Atlassian Is Building AI for Real Developer Work with Kun Chen
Description
In this RedMonk Conversation, James Governor sits down with Kun Chen, Lead Principal Engineer at Atlassian, to explore what “context engineering” really means in practice—and why it matters far more than one-shot “vibe coding” demos for real software teams. Drawing on Atlassian’s internal experience building Rovo Dev, Kun explains how AI agents become genuinely useful when they are grounded in the lived context of professional development: Jira history, pull requests, CI/CD pipelines, incidents, and organizational standards. The discussion covers how Rovo Dev reduces developer toil, integrates directly into existing workflows and leverages emerging standards such as MCP to stay extensible. Along the way, they dig into hard engineering tradeoffs around models, tokens, cost, and signal-to-noise in code review, and look ahead to a future where AI agents take on more autonomy—shifting the bottleneck from writing code to deciding what’s worth building.
This RedMonk conversation is sponsored by Atlassian.
United States
