SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Cognition AI is building Devin, the first AI software engineer and one of the most capable autonomous agents in production. Devin writes and runs code, uses tools, takes actions in real customer systems, and operates for hours without human oversight. This makes safety and alignment critical—and one of the few places where safety work ships to millions of developers rather than remaining theoretical.
You will build Cognition's safety function from the ground up. This is a hands-on, high-impact role reporting to leadership, responsible for the research agenda, evaluation and red-teaming program, deployment policies, and team building.
Key responsibilities:
- Own safety end-to-end: Set the safety strategy for Devin and its underlying models, and own the outcomes.
- Build the evaluation program: Design evals and red-teaming for agentic risks including unsafe actions, prompt injection, data exfiltration, sandbox escape, reward hacking, and misuse. Integrate these into every model and product release.
- Shape training and agent harness: Partner with pre-training, post-training, and agent teams to embed safety into model training and how Devin plans, acts, and requests help.
- Set deployment policy: Define what Devin can and cannot do, how permissions and oversight work, and navigate gray areas in a way that satisfies enterprise customers.
- Build and lead the team: Hire and lead safety researchers and engineers. Represent Cognition's safety work externally with customers, policymakers, and the research community.
Cognition's founding team includes world-class competitive programmers, former founders, and leaders from Scale AI, Palantir, Cursor, Waymo, Tesla, Lunchclub, Modal, Google DeepMind, and Nuro. The company was founded in 2023 and is headquartered in San Francisco.
Requirements:
- Frontier lab safety or alignment experience: Hands-on work on safety, alignment, evaluations, or red-teaming at a frontier AI lab. You must have shipped safety work that affected real models or products.
- Agentic systems depth: Understanding of how agents fail in practice (tool misuse, specification gaming, long-horizon drift, adversarial inputs) and ideas for measuring and mitigating these failures.
- Research credibility: Track record of published or widely used safety research, evals, or methods. Advanced degree in Computer Science, Machine Learning, or related field is a plus.
- Technical fluency: Comfortable in Python and the training/inference stack. Able to read code, run evals, and engage deeply with researchers on technical details.
- Builder mindset: Preference for starting or scaling safety functions and shipping mitigations over writing memos.
- Judgment under uncertainty: Ability to make clear calls on deployment risk with incomplete information and explain them to engineers, executives, and customers.
About Cognition AI
AI / Data / Infrastructure — AI software engineering agents and developer automation.
