Research Engineer, Safety & Alignment
Core
Build evaluations, red-teaming, and alignment methods for autonomous AI agents to ensure safe production behavior.
Role type
Senior IC research engineer (safety & alignment)
Builds
Autonomous AI agents (Devin)
Domain
Applied AI / Frontier AI
Deliverable
production ML models
Required skills
Safety or alignment research experience, Agentic systems understanding, Python, PyTorch (or JAX), Empirical rigor, Adversarial thinking
Preferred skills
Experience at a frontier AI lab, Applied AI company, or developer tools company
Technologies
Python, PyTorch, JAX
Responsibilities
Design and run evaluations for failure modes (unsafe actions, prompt injection, data exfiltration, sandbox escape, reward hacking), Red-team the agent and harness, Develop alignment methods (reward modeling, preference data, constitutional approaches), Partner with agent team on permissions and oversight, Publish safety work
Seniority
Senior, hands-on IC
