Post-Training ML Research Engineer - LLMs & RL/SFT

Google; Town of Montana (WI); On-site; USD 174,000 - 252,000; Full time

Job summary

Google DeepMind seeks a Research Engineer to advance Gemini code post-training using reinforcement learning and supervised fine-tuning for large language models. You will develop scalable post-training pipelines, collaborate with Operations Research teams to benchmark performance, and design automated evaluation suites to ensure robust improvements.

You will contribute to infrastructure, reward models, and data curation for iterative model releases while engaging with a global research ecosystem

Qualifications

• Bachelor's degree in CS/ML/AI or equivalent practical experience.
• 5 years of experience developing ML models with JAX, PyTorch, or TensorFlow.
• Experience with post-training RL or supervised fine-tuning for LLMs.
• Experience designing or running LLM evaluation benchmarks and pipelines.

Responsibilities

• Drive post-training research and engineering using RL and SFT to advance Gemini coding capabilities.
• Develop and scale post-training pipelines and benchmarks with OR teams to set industry-leading performance.
• Design, build, and maintain evaluation suites and automated benchmarks.
• Implement training infrastructure, reward models, and data curation workflows.

Skills

ML research

Frameworks: JAX/PyTorch/TensorFlow

Post-training RL / SFT for LLMs

LLM evaluation benchmarks

Education

Bachelor's degree in CS/ML/AI or equivalent

Tools

TPU/GPUs acceleration

Job description

Google DeepMind seeks a Research Engineer to advance Gemini code post-training using reinforcement learning and supervised fine-tuning for large language models. You will develop scalable post-training pipelines, collaborate with Operations Research teams to benchmark performance, and design automated evaluation suites to ensure robust improvements.

You will contribute to infrastructure, reward models, and data curation for iterative model releases while engaging with a global research ecosystem