How Meta Scaling AI-Native Engineering Tripled Code Coverage (And Cut Slop)

Meta Software Engineer Ian Thomas presents how Reality Labs scaled an AI-for-productivity community from 0 to 400+ engineers with 80% tool adoption, using a DORA-based AI maturity framework with 6 integration dimensions. The video showcases an unsupervised agentic workflow achieving 90% test coverage in just 3 hours of manual effort, plus case studies on Unity code migrations and MCP queries. Targeted at senior developers, architects, and engineering leaders seeking battle-tested frameworks for AI-native workflows without sacrificing code quality.