JOBSEARCHER

Research Engineer, Gemini Code Post-training, DeepMind

ARCHIVED
GoogleSan Jose, CAL6 LeadSeptember 15th, 2026

We can't find an active application page for this role right now. It may reopen or be listed elsewhere. Use Next Steps to search for an active apply link and similar live jobs.

Overview: In this role, you drive post-training research and engineering to advance Gemini-style coding capabilities using RL and SFT. You will build scalable agentic pipelines and benchmark suites in collaboration with OR teams to achieve top-performing Code Arena results. You design evaluation frameworks and automation to stress-test and improve model performance, and you implement training infrastructure and data curation to accelerate iterative releases. You contribute to a fast-paced research environment with broad impact across products and partners. Compensation / Benefitsbonus target (15%)equitybenefits package
Responsibilities: Lead post-training RL and SFT workflows for large language modelsDesign and maintain frontier evaluation suites and automated benchmarksBuild and scale training infrastructure, reward models, and data curation pipelinesCollaborate with Operations Research teams to establish top benchmark performanceDevelop pipelines to accelerate iterative model improvements from revision to revisionContribute to cross-functional projects enabling large-scale experiments and rapid deploymentSupport research-to-product transitions by aligning experiments with real-world use casesEngage in publication and research community activities when possible
Key requirements: Bachelor in Computer Science, Machine Learning, Artificial Intelligence, or related field; or equivalent practical experience 5 years of software engineering or research experience developing ML models with JAX, PyTorch, or TensorFlowExperience conducting post-training RL or SFT for LLMsExperience designing or running LLM evaluation benchmarks and measurement pipelinescollaborativefast-pacedambitious problem-solverJAXPyTorchTensorFlow