Ep 10 – Multi-Agent Engineering: Building Disagreeing Systems for Higher Quality | Devashish Meena скачать в хорошем качестве

Ep 10 – Multi-Agent Engineering: Building Disagreeing Systems for Higher Quality | Devashish Meena

AI is making engineers faster. But is it making teams better? In Episode 10 of The Agentic Quality Podcast, Tatyana Arbouzova speaks with Devashish Meena, Senior Director of Engineering at Esper, about building multi-agent systems that improve code quality through structured disagreement. Devashish shares how his team: • Created AI agents that mimic senior engineers • Designed “anti-agents” that intentionally disagree • Broke large pull requests into smaller deployable units • Codified tribal knowledge into structured AI context • Built pre-review AI systems before human code review This episode explores review fatigue, context fragmentation, AI over-agreeability, and why generating more code is the wrong metric. The future of engineering is not faster typing. It is engineers evolving into agent orchestrators. Explore more episodes - https://contextqa.com/blog/ #AgenticQuality #AIEngineering #MultiAgentSystems #CodeQuality #ContextQA