RD-Agent
by microsoft
RD-Agent is a Microsoft framework that uses large language models to automate data and model research and development, such as machine learning engineering, Kaggle-style competitions and quantitative finance strategies. It proposes ideas, writes and runs code, and learns from the results in a loop.
- Stars
- 15.2k +18 a day
- Licence
- MIT
- Runs on
- Your computer
- Setup
- Technical
- Language
- Python
- Age
- 944 days
What you can do with it
- Run an automated data science agent on Kaggle-style machine learning tasks
- Automate the search for and testing of quantitative finance factors and models
- Fine-tune language models using an agent-driven workflow and view traces in a web interface
Make money with it
A freelance data scientist could use it to speed up delivery of prediction models or analysis for clients.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- A small data team could automate trial-and-error model building and feature engineering, saving analyst hours.
- A freelancer offering machine learning work could use it to produce baseline models faster for client projects.
Can you use it commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
It is a Python package installed from PyPI (rdagent) and configured with an LLM provider through LiteLLM; the documentation covers the scenario-specific setup, which may involve Docker.
Follow the official instructions in the project README.
Things to check
- Needs paid LLM API keys, and long agent runs can be costly
- Some scenarios such as model fine-tuning probably need a GPU
- The quantitative trading scenario is research-oriented and carries financial risk if used for real trading
- Setup is research-grade and may be complex
Checked 03 Nov 2026. Summary drafted by AI from the project's README.
