I’m an engineer at Amazon building AI-native development tooling to accelerate work in Customer Service. I spend most of my time at the intersection of AI agents and developer productivity — figuring out how to make coding agents actually useful in day-to-day engineering work, not just demos.

This blog is where I share techniques, tools, and lessons from that work. Everything from agent memory systems and deterministic guardrails to multi-agent orchestration and session-to-session learning.

I maintain a collection of open-source extensions for the pi coding agent:

• pi-memory — persistent memory that learns from sessions
• pi-steering-hooks — deterministic tool-call guardrails
• pi-ralph — multi-agent orchestration loops
• pi-lsp-extension — LSP integration for code intelligence
• pi-knowledge-search — semantic search over local files
• pi-session-search — search past coding sessions

Find me on GitHub.