General Analysis
Research Engineer, Post-Training
San Francisco, CA · Full-time · Posted Oct 29, 2026
About the job
Post-train models for adversarial capabilities using reinforcement learning, building the training pipeline from data generation and environment design through evaluation.
What you'll do
- Build and operate training pipelines, including data generation, environment design, and evaluations.
- Address distributed-systems challenges and develop methods for data-constrained, low-signal-to-noise settings.
- Improve algorithms for applying trained models to adversarial simulations.
What they ask for
- Comfort with PyTorch, RL infrastructure, post-training infrastructure, or tools such as Tinker and verl.
- Academic and/or industry experience with reinforcement learning or machine learning research, or eagerness to learn.
Nice to have
- Publications in top-tier venues such as NeurIPS, ICML, ICLR, CVPR, ACL, or similar.
- Security experience, or enthusiasm for building it.
Tools
PyTorch, Tinker, verl
AI focus
Uses reinforcement learning to post-train models for adversarial capabilities and simulations, with work spanning training infrastructure and evaluation.
Model-written summary · the original posting is authoritative