An autonomous ML research assistant that investigates, writes, and ships production-ready ML code using the Hugging Face ecosystem — with deep access to docs, papers, datasets, and cloud compute.
Searches HF docs, papers, datasets, and model repos on your behalf. No more context switching.
Writes and runs code in isolated Modal sandboxes. No local setup required. A10G, H100 — you name it.
From research to fine-tuning to deployment. Pushes trained models and results directly to HF Hub.
React UI streams every tool call, approval gate, and execution result in real-time.