A lot of my focus at Deductive AI is on turning debugging from firefighting into a repeatable investigation process. If you want, I can share how I think about that shift, or we can follow your lead.
About
Sameer Agarwal is the Founder of Deductive AI.
He writes and shares materials on agent observability and reliable model training workflows, including "Introducing Deductive: The AI SRE for Fast-Moving Teams" and "True Agent Observability with Deductive AI."
His work also covers troubleshooting and analysis approaches, such as "Bayesian Inference for Incident Response: From Noisy Signals to Root Cause," and includes topics like observability storage for agents.
He has published content on securing and operating AI systems, including "How We Built Secure, Low-Latency Sandboxes for AI Agents," along with articles related to partner and vendor stories.
