Principal Applied Scientist (User Modeling)

What you’d do

Own LLM post-training pipelines including SFT, DPO, and reinforcement fine-tuning on GPU infrastructure, build reward models, and design evaluation using LLM-as-a-Judge. Full-time remote role within the US, with technical leadership and mentorship duties. Found on the employer's careers page on Oct 20, 2026. Apply through the link.

You’ll need

• LLM post-training: SFT, DPO, RFT/GRPO
• GPU infrastructure training experience
• Reward model (PRM) development
• LLM-as-a-Judge evaluation frameworks
• Technical leadership and mentorship