Lead RL Research Scientist – Post-Training for LLMs & Code

AMD; Santa Clara (CA); On-site; USD 180,000 - 300,000; Full time

Job summary

AMD, a leading semiconductor innovator in Santa Clara, seeks a Lead AI Research Scientist to advance reinforcement learning for post-training LLMs and code models. You will design reward models, curricula, and training recipes, addressing sparse or expensive labels from experts and simulators, while pushing robust, production-ready RL methods.

You will publish results at premier conferences and collaborate with infra to scale training, contributing to a practical RL roadmap that balances

Qualifications

• PhD preferred with strong publication record.
• Experience training RL or preference-based models at scale.
• Experience with LLM post-training, RLHF/RLAIF, or policy optimization.

Responsibilities

• Research and develop RL methods for post-training LLMs and code models on structured engineering tasks with verifiable or preference-based feedback.
• Design reward models, curricula, and off-policy or on-policy training recipes suited to sparse, noisy, or expensive labels from experts and simulators.
• Characterize failure modes (reward hacking, degenerate policies, instability) and propose mitigations grounded in experiments.
• Collaborate with RL infra engineers to scale training; define interfaces for rollout generation, logging, and reproducibility.
• Publish at top venues (e.g. NeurIPS, ICML, ICLR) and contribute internal technical leadership on the RL roadmap.

Skills

RL theory

RL algorithms

Academic publishing

GPUs

Distributed training

Code models

LLMs

RLHF/RLAIF

Policy optimization

Large-scale codebases

Education

PhD in Computer Science or ML

Tools

PyTorch

TensorFlow

JAX

MPI

Ray

Job description

AMD, a leading semiconductor innovator in Santa Clara, seeks a Lead AI Research Scientist to advance reinforcement learning for post-training LLMs and code models. You will design reward models, curricula, and training recipes, addressing sparse or expensive labels from experts and simulators, while pushing robust, production-ready RL methods.

You will publish results at premier conferences and collaborate with infra to scale training, contributing to a practical RL roadmap that balances