A selection of my most recently active open source work. Most of it lives on GitHub.
Machine learning
mlx-ruby — Ruby bindings for Apple’s MLX array computing and machine learning framework. Brings high-performance ML on Apple Silicon to Ruby, covering the full MLX API including array operations, automatic differentiation, and neural network building blocks.
mlx-ruby-lm — An LLM inference toolkit built on top of mlx-ruby. Provides CLI tools and Ruby APIs for running language models locally on Apple Silicon hardware.
mlx-onnx — A standalone C++ and Python library for exporting MLX compute graphs to the ONNX format, enabling cross-platform model portability from MLX to other runtimes.
mlx-ruby-examples — Example projects and notebooks demonstrating practical usage of the mlx-ruby library.
Hardware and compilers
rhdl — A hardware description language and simulator written in Ruby. Design digital circuits in a Ruby DSL, simulate them, synthesize to gate-level netlists, and export to Verilog. Includes browser-based simulation and example designs for the MOS 6502, Apple II, Game Boy, and RISC-V.
Developer tools
herd — A Rust-based monitoring and control tool for steering Codex and Claude agent sessions running in tmux. Offers both a CLI and an interactive TUI for managing multiple concurrent AI coding sessions.
deep-research — A collection of structured deep dives on technical topics spanning AI tooling, hardware design, and systems programming.