Show HN: Gittins, an online bandit engine that enforces log correctness

Category: ai-ml

Tags: contextual-bandits, reinforcement-learning, online-learning, rust, python-bindings, wasm

Score: 6.8/10 (Innovation: 6, Technical: 8, Documentation: 7, Utility: 6)

Gittins is a production-grade contextual bandit engine with a Rust core and Python/WASM bindings, emphasizing online learning, safe reward handling, and deterministic performance. It uniquely enforces log correctness, supports dynamic actions, and provides offline policy evaluation, making it suitable for real-world deployment. The project stands out for its practical design choices and performance optimizations.

Target audience: data scientists, machine learning engineers, backend developers

Repository: https://www.getgittins.dev · Python · MIT

View on Hacker News