What it does
R&D-Agent from Microsoft aims to automate research and development work focused on data and AI models. The AI proposes ideas, writes code to try them, checks the results and learns from them in a loop.
It includes scenarios for data science, Kaggle competitions, finding signals for stock trading, and fine-tuning language models. Its makers report strong results on MLE-bench, a test of AI agents on machine-learning engineering tasks.
It has a web interface for watching the agent's work in real time.
Who it's for
Data scientists and AI researchers who want to automate experiments.
Can I use it without coding?
Developers only. Needs technical setup and keys for AI services. New to this? Start here.
Can I run it?
- Cost
- Free, but needs an AI accountThe software is free, but it uses an online AI service (like OpenAI or Anthropic), which usually charges for use.
- Runs on
- Linux
- API key
- Needed. An API key is a password-like code from an online AI service. What's that?
Try it
- Watch the demos on the project's live demo website to see what it does.
- To run it yourself, you need a Linux computer and an account with an AI provider.
Using Claude Code, Cursor or another AI coding assistant? Paste the prompt and it will walk you through the setup.New to this?
Before you install
- Only download from the official GitHub page. Copies with similar names exist for many popular projects.
- License (MIT): Very permissive: free to use, change and share, including for business.
- Keep your API key private. Never paste it into websites or share it, and set a spending limit with the AI service.
- More on staying safe
Where people are talking about it
- Agent² RL-BenchResearch paper · Apr 2026
- FT-Dojo: Towards Autonomous LLM Fine-Tuning with Language AgentsResearch paper · Mar 2026
- Reasoning as GradientResearch paper · Mar 2026
- R&D-Agent technical reportResearch paper · May 2025
- R&D-Agent-QuantResearch paper · May 2025