Hugging Face Runs an Overnight Agent to Rescue Open-Source Models Buried on arXiv
NielsRogge · x · 2026-08-24
Hugging Face ML engineer Niels Rogge built an overnight agent (on HF Jobs) that scans new arXiv papers, finds open-source models/datasets buried in GitHub repos or Drive links, and helps authors publish them to the Hub — a task he previously did by hand until it outgrew him.
Using Langfuse traces grouped by arXiv ID, he could inspect every prompt, output, and extraction step in one place. The visibility exposed an unreliable classifier and missing GitHub README context; after two rounds of error analysis and fixes, the agent drew only two negative comments across thousands of issues and PRs.
Related event: Hugging Face Uses Agents for Automation(3 posts)→
More from coding & agent
- Vercel Goes Full Backend: Python GA, Containers, and Daemons — cramforce · 2026-10-16
- WorldMemArena Accepted at NeurIPS: Memory Storage ≠ Memory Use — xwang_lk · 2026-10-16
- Vercel Ship Demos Container Deploys for Go, Ruby, Java and More — edwin · 2026-10-16
- TRL adds day-0 support for fine-tuning Muse Spark Open's 816B MoE with LoRA on 64 H100s — QGallouedec · 2026-10-16
- Databricks engineers land impactful PRs on day 2 with AI dev tools, vs week 3 before — Yuchenj_UW · 2026-10-16
- Perplexity open-sources SPACE, the sandbox platform powering its Computer agents — AravSrinivas · 2026-10-16