How I automate my own job at Hugging Face using agents — Niels Rogge, Hugging Face
Summary
Hugging Face ML engineer automated his entire job using AI agents to scale research artifact discovery from manual GitHub issues to hundreds of papers daily. By building a deterministic workflow agent instead of a fully autonomous one, he solved the critical problem of researchers publishing models on third-party services rather than centralized platforms.
Key Takeaways
- Identified a scalability bottleneck: manually opening GitHub issues for hundreds of daily research papers became unsustainable, requiring agent-based automation to maintain community outreach.
- Created a deterministic workflow agent rather than autonomous agent for predictability and control—reflecting Anthropic's 2024 guidance that workflow-based approaches are more reliable for production use cases.
- Solved researcher discoverability problem by centralizing ML artifacts on Hugging Face with metadata tags and model cards, enabling filtering by task type, language, and library compatibility.
- Developed automated workflow that replicates manual process: find GitHub URL → read README → check if artifacts exist on Hugging Face → open issues or PRs → follow up with authors.
- Applied win-win business logic: centralizing research artifacts improves discoverability for users AND provides documentation/tooling benefits plus promotional reach for researchers publishing work.
Related topics
Transcript Excerpt
[music] >> Okay. All right. Hello everyone. Thanks for coming by. Today I'll talk about how I automate my own job at Hugging Face using agents. Um short introduction. I'm just uh Niels from Belgium, the land of beer, fries, and chocolate. I studied at KU Leuven, and I'm a machine learning engineer at Hugging Face for 5 years now. Uh today I'll talk about the community science team at Hugging Face, which is the team I'm part of. Uh then I'll talk about how I automate large parts of the community science team. And finally, I'll also discuss some other efforts uh that we do at Hugging Face. So, let's start with the community science team at Hugging Face. So, basically, this started when I was sent I was seeing like trending research uh passing by on GitHub. And a lot of times when I saw new i…
More from ai.engineer
- The Era of Compound Engineering — Kieran Klaassen, Every/Cora
- Agentic SDLC at Uber — Uday Kiran Medisetty & Adam Huda, Uber
- LLM Knowledge Bases: a practical guide — Ben Holmes, Warp
- The Dark Arts of Web Automation: Teaching Agents to Use Websites Like Humans — Corey Gallon, Rexmore
- Evolution of agentic surfaces — Gagan Bhat & Isabella Kai He, Anthropic