Member of Technical Staff (Post Training)
Core
Lead post-training of state-of-the-art foundation models for open-ended agentic capabilities in scientific research, designing algorithms and pipelines for autonomous research.
Role type
Senior IC deep learning researcher (post-training)
Builds
Post-training pipelines, autocurricula, judges, harnesses, and eval systems for foundation models
Domain
AI research, recursive self-improvement, scientific discovery
Deliverable
production ML models
Required skills
deep learning research, post-training LLMs/vision/video/multi-modal models, SFT and RL algorithms, Python, deep learning frameworks (PyTorch/JAX), coding agents, experimental design
Preferred skills
PhD in math/CS/hard science, RL at scale (GRPO/PPO/DPO/distillation), distributed/long-context training infrastructure, autocurricula/meta-learning/recursive self-improvement, frontier model post-training experience
Technologies
PyTorch, JAX, coding agents
Responsibilities
Design and tune SFT/RL algorithms for autonomous research models; Build autocurricula, judges, and eval pipelines; Run large-scale experiments on SOTA hardware; Close recursive loops for AI agent self-improvement; Collaborate on hardware optimization for scientific domains
Seniority
Senior, hands-on IC