I follow my instincts, especially when a system feels one abstraction away from being interesting. I like figuring out how things behave under load, across unreliable networks, and when the happy path quietly disappears.
- I build with Rust, TypeScript, Node.js, and Python - mostly backend, systems, and developer-tool territory.
- I enjoy concurrency, caching, distributed systems, APIs, and making failure modes less mysterious.
- I like to manipulate LLMs: give them strange constraints, prod the edges of their reasoning, and occasionally make them earn their tokens.
- LeetCode Knight (1924, top 3.8%) with 600+ problems solved.
01 Design for failure: retries, clear errors, observable systems.
02 Make the common path fast: cache deliberately and measure the result.
03 Keep interfaces boring: well-defined APIs outlive clever implementations.
04 Follow the hunch - then test it hard enough to trust it.
If you are working on backend infrastructure, developer tools, systems software, or applied ML, I would be glad to connect. The fastest way to reach me is email.
