Sylvain Kalache
Sylvain Kalache is the Head of Rootly AI Labs. A former Senior SRE at LinkedIn, he co-designed a patented self-healing infrastructure. He later co-founded Holberton, training software engineers worldwide. He also writes about AI, automation, and DevOps for TechCrunch, Forbes, and The New Stack.
Session
Developers using GenAI tools ship 25% more tasks, accept a billion lines of AI-generated code per day, and are more productive than ever. So why are production incidents up 3x?
This session shares what we found after analyzing incident trends across thousands of engineering teams and why the productivity gains from AI-assisted coding are creating an invisible reliability tax that most organizations aren't yet measuring.
We'll walk through the Incident Rate Formula (λ + C × P): as AI accelerates the rate of change (C), the probability that any given change causes failure (P) becomes the only lever infra engineering teams can control. We'll look at how AI actually fails in production, echo chambers in unit tests, slop squatting attacks, hallucinated dependencies, and why the "angry intern" mental model is the most useful frame for working safely with LLMs.
The talk closes with practical strategies for keeping P low: mutation testing at scale (Meta's 70%+ acceptance rate), AI-augmented code reviews, MCP-powered incident investigation, and where AI SREs are already handling Sev-2 and Sev-3 incidents autonomously.
If you're building with AI or leading teams that are, this talk will give you a data-driven framework for staying fast without letting reliability quietly collapse.