Researcher, Post Training
Core
Own the full post-training pipeline for large language models to improve code generation and agent workloads, translating research into production recipes.
Role type
Senior IC machine-learning engineer (post-training)
Builds
Production-ready LLMs for code generation and autonomous agents
Domain
Generative AI, Large Language Models, Software Development
Deliverable
production ML models
Required skills
post-training (RFT/RLVR/preference optimization), distributed training, GPU cluster management, PyTorch or JAX, evaluation system design, model quality regression analysis
Preferred skills
code generation or agentic use cases, speculative decoding, open-source ML contributions
Technologies
PyTorch, JAX, GPU clusters
Responsibilities
Own the full lifecycle of post-training from data curation to deployment; apply reinforcement learning and preference optimization methods; build evaluation infrastructure; operate production training systems; resolve end-to-end failures; read papers and run experiments
Seniority
Senior, hands-on IC
