Researcher, Post Training

Skills

About the Role

Own the full post-training lifecycle, from data curation and training jobs through evaluation and deployment. Apply reinforcement learning, preference optimization, and supervised fine-tuning to improve code generation, user-intent reasoning, and agent reliability while building evaluation infrastructure and operating production-scale training systems.

Requirements

• Have personally run post-training jobs on large language models using RFT/RLVR, preference optimization, or similar methods
• Write solid production code
• Be fluent in PyTorch or JAX
• Be comfortable with distributed training setups and GPU clusters
• Understand preference optimization, reward modeling, and alignment techniques
• Have built or significantly contributed to evaluation systems that capture real-world quality
• Be able to trace model-quality regressions through serving, inference, and training

Responsibilities

• Own the full post-training lifecycle from data curation and training runs through evaluation and deployment
• Apply and adapt reinforcement learning, preference optimization, and supervised fine-tuning methods
• Build evaluation and experimentation infrastructure covering helpfulness, safety, latency, and reliability
• Develop and operate production systems for large-scale training jobs, including GPU orchestration and data pipelines
• Work with agent, product, and infrastructure engineers to turn model gains into product improvements
• Investigate and resolve failures across training recipes, data, and serving
• Read papers, run experiments, and move promising research into production quickly