SOFTWARE ENGINEER · AI SYSTEMS
LLM-powered products from prototype to production — autonomous agents that plan, execute, and self-correct against measurable rubrics.
A few things I've designed, built, and operated.
Autonomous agents pull tickets from a board, run their own tools in a managed sandbox, self-grade, and report scores back to the card. Built on Anthropic Managed Agents.
P95 latency observability for SQL-heavy backends. Pinpoints the query stealing your budget — without sampling overhead.
Zero-downtime Postgres schema migrations with rollback paths derived automatically from the DDL graph.
I've spent the last decade shipping production software across fintech, developer tools, and AI infrastructure. My current focus is reliable multi-agent orchestration: giving an LLM enough structure to make real progress while keeping a human in the loop where it matters.
I care about latency, observability, and making the boring parts of software boring on purpose. Comfortable across the stack; happiest at the API design and infrastructure layer.
Open to consulting on agentic systems and AI infrastructure.
burak@example.com