Research Engineer, Gemini Cyber Post-Training, DeepMind
Mountain View, CA, USA +1 · Full-time · Posted Oct 22, 2026
- Location
- Mountain View, CA, USA
New York, NY, USA - Experience
- 2 years of industry or academic experience working in Reinforcement Learning environments.
- Education
- PhD degree in Computer Science, or a related field, or equivalent practical experience.
About the job
Develop cybersecurity models and scalable infrastructure for Gemini, including vulnerability-focused post-training, agent evaluations, and integration of vulnerability discovery and patching capabilities.
What you'll do
- Build RL environments and supervised fine-tuning datasets; iterate on post-training recipes and analyze training runs.
- Build evaluation pipelines, benchmarks, and leaderboards to measure agent performance.
- Contribute to Cyber Strike’s finding, fixing, and exploit tracks, including dataset curation and CodeMender integration.
What they ask for
- PhD in Computer Science or a related field, or equivalent practical experience.
- 2 years of industry or academic experience working in Reinforcement Learning environments.
- Experience building distributed system infrastructure.
- At least one publication record in vulnerability research, such as auditing, reverse engineering, or exploitation.
Nice to have
- Experience with program analysis techniques, such as static analysis, symbolic execution, or fuzzing.
- Background in machine learning and statistics.
Pay
$147000 - $210000 USD base salary
Tools
Gemini, CodeMender, Reinforcement Learning, RL environments, SFT
AI focus
Develop cybersecurity models and post-training methods for vulnerability discovery; build agentic evaluations and integrate vulnerability discovery and patching capabilities.
Model-written summary · the original posting is authoritative