Autonomous Agent Safety & Alignment Engineer

Cognition; San Francisco (CA); On-site; USD 210,000 - 260,000; Full time

Job summary

Cognition is building Devin, an AI software engineer. We seek a founding member of the Safety team to design evaluations, perform red‑team testing, and develop alignment methods that shape how models are trained and deployed.

You will work with the agent team to define when the agent should act, when it should ask for help, and how it should explain itself, with work published as papers and open methods where appropriate.

Qualifications

• PhD in Computer Science, ML, Physics, Mathematics, or related discipline; or equivalent industry research experience.
• Hands-on safety or alignment research experience in frontier AI labs or industry.
• Experience with evaluations, red-teaming, alignment, or interpretability.
• Strong Python programming and reliable engineering skills; ability to build eval infra and run experiments at scale.
• Familiarity with agentic systems, tool use, and adversarial thinking.

Responsibilities

• Build agentic safety evals: Design and run evaluations for the failure modes that matter for autonomous agents: unsafe actions, prompt injection, data exfiltration, sandbox escape, reward hacking, and misuse. Make them fast enough to run on every model and product release.
• Red-team Devin: Attack the agent and its harness systematically. Find the failures before customers do and turn them into fixes and regression tests.
• Develop alignment methods: Work with post-training on reward modeling, preference data, constitutional approaches, and other techniques that make the model safer without making it worse at the job.
• Shape the agent harness: Partner with the agent team on permissions, oversight, and escalation: when Devin should act, when it should ask, and how it should explain itself.
• Publish and share: Contribute to Cognition's external safety work through papers, evals, and open methods where it makes sense.

Skills

Python

Engineering fundamentals

Education

PhD in Computer Science, ML, Physics, Mathematics, or related discipline; or equivalent industry research experience

Tools

PyTorch

JAX

Job description

Cognition is building Devin, an AI software engineer. We seek a founding member of the Safety team to design evaluations, perform red‑team testing, and develop alignment methods that shape how models are trained and deployed.

You will work with the agent team to define when the agent should act, when it should ask for help, and how it should explain itself, with work published as papers and open methods where appropriate.