2025-08-26 –, Potter Auditorium - Kenneth C Rowe Management Building
Automated incident resolution isn’t new. But in 2025, AI-powered tools are making promises that almost sound too good to be true. The key question is: Can AI really be trusted with production systems?
Automated incident resolution isn’t new—Facebook’s FBAR and Dropbox’s Naoru pioneered self-healing infrastructure over a decade ago. Today, AI-powered tools promise faster root cause analysis and remediation, reducing MTTR dramatically. But can AI be trusted with production systems?
In this talk, Sylvain, a former SRE at LinkedIn, explores the history of self-healing infrastructure (a topic in which he holds a patent), the evolution of AI-driven RCA, and practical strategies for adopting these tools safely. Attendees will gain insights into leveraging AI for reliability while avoiding common pitfalls—ensuring that automation enhances rather than replaces human expertise in modern SRE workflows.
Sylvain is an entrepreneur and software engineer. He holds a patent on self-healing architectures, and is currently leading the AI Labs at Rootly, a research group dedicated to pushing the boundaries of LLMs and ML in incident response. His education company, Holberton, has trained thousands of students who have landed FAANG jobs in over 25 countries.