Discover a comprehensive framework for building high-quality AI agents that serve as expert partners for software engineers in this 41-minute conference talk. Learn to prevent low-quality AI output while creating context-aware assistants that analyze large codebases, evaluate architectural alternatives, and reason about complex datasets. Explore practical strategies for establishing evaluation frameworks to compare AI developer tools and develop specialized workflows for building agents focused on performance analysis and architecture validation. Gain actionable insights for empowering engineering teams with AI assistants that provide meaningful support and context during daily development tasks, ultimately enhancing code quality and accelerating development processes.