Member of Technical Staff - Post-Training Research
Core
Conducting hands-on post-training research for Large Language Models (LLMs) to optimize training, deployment, and production observation, bridging research bets with production workloads.
Role type
Post-training research scientist (LLM infrastructure)
Builds
Production-ready post-training frameworks, distributed-training approaches, and online training for deployed models.
Domain
Artificial Intelligence / Machine Learning / Cloud Infrastructure
Deliverable
production ML models
Required skills
Post-training LLM research, RL (async/agentic/on-policy), model distillation, long-context RL, routing models, product sense for frontier techniques, shipping research to production, cross-functional collaboration with engineering and customers.
Preferred skills
Experience with speculators, KV injection, blockwise parallel drafting, DiLoCo, evolutionary strategies.
Technologies
LLMs, RL, distributed training, GPU sandboxes, production traffic data.
Responsibilities
Own end-to-end post-training research bets, work with customers to train models and integrate learnings, collaborate with external research labs, work with engineering to turn techniques into products, help shape the research agenda.
Seniority
Mid-Senior, hands-on IC