Head of Safety

The role

You will build the safety function from the ground up, including the research agenda, evaluation and red-teaming program, deployment policies, and the team. You will work with researchers training models and engineers building the agent harness.

What you will do

• Set the safety strategy for Devin and the models behind it.
• Design evals and red-teaming for agentic risk.
• Make evals part of every model and product release.
• Partner with pre-training, post-training, and agent teams.
• Define deployment policy and how permissions and oversight work.
• Hire and lead safety researchers and engineers.
• Represent Cognition's safety work with customers, policymakers, and the research community.

Required qualifications

• Hands-on work on safety, alignment, evaluations, or red-teaming at a frontier AI lab.
• Shipped safety work that affected real models or products.
• Understanding of how agents fail in practice including tool misuse, specification gaming, long-horizon drift, and adversarial inputs.
• Track record of published or widely used safety research, evals, or methods.
• Comfortable in Python and in the training and inference stack.
• Ability to read code, run evals, and argue with researchers on details.
• Experience starting or scaling a safety function.
• Ability to make clear calls on deployment risk with incomplete information.

Nice to have

• An advanced degree in Computer Science, Machine Learning, or a related field

About the company

Cognition is an applied AI lab building end-to-end software agents. They are the makers of Devin, the first AI software engineer.