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 agent. The company is seeking a Research Engineer for Safety & Alignment to join as a founding member of the safety team, reporting to the Head of Safety.
In this hands-on research engineering role, you will work on real-world safety challenges for an agent that writes code, runs tools, takes actions in customer systems, and operates autonomously for hours without human intervention. Safety here means ensuring the agent behaves correctly in production millions of times daily—not just publishing research papers.
Key responsibilities include:
- Building agentic safety evaluations: Design and implement evaluations for failure modes critical to autonomous agents, including unsafe actions, prompt injection, data exfiltration, sandbox escape, reward hacking, and misuse. Create evaluations fast enough to run on every model and product release.
- Red-teaming Devin: Systematically attack the agent and its harness to find failures before customers encounter them, then convert findings into fixes and regression tests.
- Developing alignment methods: Collaborate with post-training teams on reward modeling, preference data, constitutional approaches, and other techniques that improve safety without degrading agent performance.
- Shaping the agent harness: Partner with the agent team on permissions, oversight, and escalation mechanisms—determining when Devin should act autonomously, when it should request approval, and how it should explain its reasoning.
- Publishing and sharing: Contribute to Cognition's external safety work through papers, evaluations, and open-source methods where appropriate.
The founding team includes world-class competitive programmers, former founders, and leaders from Scale AI, Palantir, Cursor, Waymo, Tesla, Google DeepMind, and other cutting-edge AI companies.
Requirements:
- Hands-on safety or alignment research experience: demonstrated work on evaluations, red-teaming, alignment, or interpretability at a frontier AI lab or in published research; evidence of building systems that changed how a model was trained or deployed.
- Agentic systems understanding: knowledge of how agents fail in practice (tool misuse, specification gaming, long-horizon drift, adversarial inputs); experience measuring these failures or concrete ideas for how to measure them.
- Strong engineering fundamentals: proficiency in Python and PyTorch (or JAX); ability to build evaluation infrastructure, run experiments at scale, and read training and inference code.
- Empirical rigor: ability to design clean experiments, report results honestly, and distinguish real improvements from noise.
- Comfort with adversarial thinking: demonstrated ability and enjoyment in breaking systems.
- Relevant industry experience: prior work at a frontier AI lab, applied AI company, or developer tools company preferred.
- Advanced degree: PhD in Computer Science, Machine Learning, Physics, Mathematics, or related discipline; or equivalent industry research experience.
About Cognition AI
AI / Data / Infrastructure — AI software engineering agents and developer automation.
