I'm a machine learning engineer with a strong data science background (4 years professional ML, 7+ years Python). These days, I spend most of my time writing Rust. I've built several Rust projects focused on real systems, and real problems.
- is-it-slop: Detect AI generated slop using classic ML. Blazingly fast embedded model infrence, served in Rust using ONNX & a custom TFIDF vectorizer.
- Heisenberg: Complete offline location enrichment tool, using the geonames database to backfill and enrich location data.
- market-making: System for ingesting live market data streams, replicating orderbooks locally, and distributing market data to message queues and time-series databases. Focused on engineering a market-making system from the ground up.
- ZK Sudoku prover: Verifying the correctness of Sudoku puzzles without ever revealing the puzzle solution itself. A simple rust base implementation of a Zero Knowledge Prover based on graph colouring.
- Typed Prompt: A simple type-safe, validated prompt management system for LLMs
- Learning how to do things I dont currently know how to do.
- Rust, async internals, atomics, locks, lower-level primitives
- Compilers & systems programming
- Building things from scratch
- Swimming, running, netball, gaming
Ideally, a role where I build fast systems in Rust.