Agi Inc
AI Researcher
San Francisco Office. Posted 15 months ago
The role
You will lead training for model families that power on-device agents. This includes pretraining recipe choices, post-training, distillation, and quantization.
What you will do
- Own one or more model capabilities end-to-end from data mixture to production on-device runtime
- Design experiments and writeups
- Lead a training workstream with a clear success metric and a checkpoint that ships
- Assist infra and product engineers by turning research wins into shipped capabilities
- Assist partnerships by communicating what is possible at the next device refresh
- Assist other researchers by reading their code and making it easier to read
Required qualifications
- Send a link to your most interesting result such as a paper, blog, model card, or GitHub with one paragraph on why it matters
- Resume
- Google Scholar or LinkedIn
Benefits
- Top-tier relocation support
- Immigration support
- Permission to publish what is safe to publish
About the company
agi-inc is a stealth team of elite founders and AI researchers with backgrounds from Stanford, OpenAI, and DeepMind. They aim to build everyday AGI through trustworthy, consumer-grade agents.