# groundhog

- **Event:** [The Recursive Agents Hackathon](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon)
- **When:** Sat, Oct 17 at 9:00 AM – 10:00 PM (PDT)
- **Where:** San Francisco, CA
- **Team:** [Sidhartha Reddy Potu](https://cerebralvalley.ai/u/sid-rp), [Raghu Vamshi Hemadri](https://cerebralvalley.ai/u/rhemadri)
- **GitHub:** https://github.com/RaghuHemadri/recursive_agents_hack
- **Demo video:** https://www.loom.com/share/bb7bc8200f54451ab04f287c05a54845
- **Gallery:** https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/103

Groundhog is a team of AI agents that uses the real Azure portal the way a person does: it looks at the screen and clicks. It runs on NVIDIA Nemotron models hosted on Crusoe. The agents get better as they work: they save lessons in a shared memory and rewrite their own instructions and tools, keeping a change only if it does better on test tasks. Risky actions like deleting or submitting need a human's approval, and a checker confirms a task is really done before it counts as finished.

## More from The Recursive Agents Hackathon

- [FruitTree.AI](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/100)
- [Blackout](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/101)
- [Turing](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/102)
- [Travel](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/104)
- [Auto-Heal](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/105)
- [HelloTiger](https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/106)

---

Markdown version of https://cerebralvalley.ai/e/crusoe-recursive-agents-hackathon/hackathon/gallery/103. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt