About

Think Reason Learn is an innovative, open-source Python library that fuses the power of large language models (LLMs) with interpretable machine learning. Developed at Vela Research in collaboration with Oxford University, it provides production-grade tools for building transparent decision-making systems—perfect for high-stakes domains like venture capital, healthcare, law, etc.

Key Features

• Explainable AI: Every prediction comes with traceable reasoning paths, rules, or cluster descriptions.

• Async-First Design: Scalable LLM interactions with synchronous wrappers for flexibility.

• Modular Algorithms: Easily extend with new models under a unified interface.

Core Algorithms

• GPTree: LLM-guided decision trees for dynamic feature generation.

• RRF (Random Rule Forest): Transparent ensembles of LLM-generated YES/NO rules.

For in-depth papers and methodology, see our Research section.