I'm a Data Scientist building open-source tools for ML engineering.
My main project. LUML is the agent-native MLOps platform: you and your coding agents build models together, agents run experiments on their own, every run is tracked, and the best models go to production on your own servers.
- Scikit-LLM: scikit-learn compatible wrappers that let you use LLMs as regular sklearn estimators.
- FNNX: a flexible format to package and deploy any ML model.
- Falcon: AutoML that trains a full pipeline in one line and exports it to ONNX.
If you find any of them useful, a star ⭐ helps others discover them.