Susheem Koul

Susheem Koul is a Software Engineer at Microsoft with 7+ years of experience in product development, currently focused on designing and building agentic systems. His work spans distributed systems, AI agents, software architecture, and the challenges of bringing AI-powered systems into production. He is particularly interested in the intersection of AI and distributed systems, exploring topics such as agent reliability, observability, token governance, and deterministic execution. He regularly speaks and writes about AI engineering, agentic systems, and software architecture. Susheem presented at the AI Engineer World's Fair 2026 in San Francisco and regularly delivers technical sessions to engineering audiences. Outside of work, he writes about the philosophy of learning, software architecture, and technology on his Substack.

Microsoft, Indiasusheemk.substack.com susheem-k susheemkoul

Debugging the Agent That Refuses to Fail the Same Way Twice

AI Coding Summit NYC

Upcoming

Debugging the Agent That Refuses to Fail the Same Way Twice

Your coding agent botched a multi file refactor once in production. You rerun the exact same prompt to debug it, and it works. The model sampled different tokens, a tool returned a different payload, the context drifted, and the failure was gone. This session shows how to capture every nondeterministic input to an agent loop into an append only execution log, replay the failed run step by step, and then turn that replay into reproducible evals so the bug never quietly returns.

artificial intelligence