LLM Inference Engine in Rust
ActiveA from-scratch inference runtime for transformer-based language models, written in Rust with a focus on memory efficiency and zero-copy tensor operations.
~/projects
Open-source experiments, tools, and libraries I'm building — mostly in Rust with a focus on language models and functional design patterns.
A from-scratch inference runtime for transformer-based language models, written in Rust with a focus on memory efficiency and zero-copy tensor operations.
A Rust library exploring algebraic data types, monadic composition, and parser combinators — drawing inspiration from Haskell's ecosystem.
A composable retrieval-augmented generation pipeline with pluggable embedding backends, vector stores, and reranking strategies.
An experimental multi-agent framework where LLM agents coordinate through message-passing and shared state without a central orchestrator.
Using Rust's type system to encode state machines at compile time — illegal transitions become type errors, not runtime panics.