Member of Technical Staff - Research Software Engineer - Safety Evaluations Infrastructure
Core
Design, build, and own secure infrastructure for running sensitive AI model evaluations in high-consequence domains like CBRN and child safety.
Role type
Senior IC research software engineer (safety infrastructure)
Builds
Sandboxed execution environments, controlled data pipelines, eval-orchestration tooling, and monitoring systems for sensitive model evaluations.
Domain
AI Safety / Security Engineering / Platform Engineering
Deliverable
production ML models | infrastructure
Required skills
Python, building sandboxed/isolated execution environments, security engineering fundamentals (RBAC, encryption, audit logging), building data pipelines for sensitive data, end-to-end ownership of ambiguous problems
Preferred skills
building evaluation/benchmarking infrastructure for ML systems, experience with LLMs or agents, familiarity with dangerous-capability domains (CBRN, cyber), familiarity with compliance frameworks for sensitive data
Technologies
Python, containerization, VM isolation, secrets management, encryption tools, audit logging systems
Responsibilities
Design and build secure, sandboxed infrastructure for sensitive model evaluations; Build controlled data pipelines and storage with least-privilege access; Partner with safety researchers to translate evaluation designs into reliable systems; Build eval-orchestration tooling for high-throughput evaluations; Develop infrastructure for measuring AI capability uplift in high-consequence domains; Implement guardrails, monitoring, and compartmentalization for sensitive work; Write production-quality Python for high-throughput data processing; Improve reliability and security posture of the evaluation platform