Research Engineer, Post-Training for Code Security Analysis

Job Description

Research Engineer at Google DeepMind working on post-training strategies (SFT, RLHF, RLAIF) to enhance Gemini’s capabilities for secure code analysis, combining ML innovation, eval, and data generation to ensure safe, powerful code-security behavior.

Responsibilities

• Design and implement advanced post-training algorithms for Gemini on code security tasks
• Develop evaluation setups and metrics for secure coding and vulnerability detection
• Generate and curate training and eval data to improve model robustness and safety
• Collaborate with researchers and engineers across DeepMind and Google on secure AI deployment

Benefits