Scaling AI Reliability: Real world lessons from Mistral AI

Join Devon Mizelle (Mistral AI), Sylvain Kalache (Rootly AI), and Giovanni Rago (Checkly) for a deep dive into the evolution of Mistral’s reliability infrastructure.

As AI companies scale extremely fast, the complexity of their infrastructure often outpaces their observability capabilities.

In this session, we’ll explore the "war stories" of scaling Mistral's platform and demonstrate how the combination of Checkly’s monitoring and Rootly’s incident response has reduced cognitive load for their engineering team.

What you'll learn:

• Infrastructure Lessons: A behind-the-scenes look at the specific tooling failures Mistral encountered while scaling.
• Detection as Code: How to use Checkly and Playwright to move beyond noisy alerts and into a stable, automated monitoring workflow.
• Automated Response: The role of Rootly in bridging the gap between detecting a failure and resolving it without manual overhead.
• Future-Proofing Ops: What the next 12 months of AI-assisted monitoring will look like, from automated detection to AI-driven monitoring decisions.