AIRLHFPyTorch
Staff Research Engineer, Post-training
Frontier AI Lab (client name withheld under NDA)
San Francisco / Remote Full-time $420-620k + equity
Own post-training for a frontier model family - RLHF, preference data and evaluation - inside a small, senior research org.
About the client
A well-funded frontier lab building general-purpose models. The post-training group is deliberately small and works directly with pre-training and product. Client name is shared with shortlisted candidates after a mutual NDA.
What you'll do
- Design and run post-training experiments across RLHF, DPO and preference-data pipelines
- Own evaluation harnesses that decide what ships
- Partner with pre-training and inference teams on data and serving trade-offs
- Mentor research engineers and set technical direction for the group
What we're looking for
- Deep experience training or fine-tuning large language models in PyTorch or JAX
- Track record of published work, shipped models, or equivalent industrial impact
- Comfort operating at large GPU scale with distributed training
- Strong engineering fundamentals - this is a hands-on role
Interview process
- 01Intro call with a Progress Lane partner
- 02Technical deep-dive with the research lead
- 03Take-home or paired research session
- 04Onsite panel and offer
Client names are withheld until a mutual NDA is in place. We never approach your current employer and never share your details without your written consent.