Head of Safety
Cognition · San Francisco
About this role
## Head of Safety ### About the role Cognition is building **end-to-end software agents** (including **Devin**, the first AI software engineer). Devin can write and run code, use tools, take actions in real customer systems, and operate for hours without a human in the loop—so getting safety right is critical. In this role, you’ll build Cognition’s **safety function from the ground up**: the research agenda, evaluation and red-teaming program, deployment policies, and the team. You’ll work closely with researchers training our models and engineers building the agent harness. ### What you’ll accomplish - **Own safety end to end:** Set the safety strategy for Devin and the models behind it, and own outcomes. - **Build the evaluation program:** Design evals and red-teaming for agentic risk, including: - unsafe actions - prompt injection - data exfiltration - sandbox escape - reward hacking - misuse Ensure these are integrated into every model and product release. - **Shape training and the agent harness:** Partner with pre-training, post-training, and agent teams so safety is built into how models are trained and how Devin plans, acts, and asks for help. - **Set deployment policy:** Define what Devin is and isn’t allowed to do, how permissions and oversight work, and how to handle gray areas—while meeting enterprise requirements. - **Build the team and external voice:** Hire and lead safety researchers and engineers; represent Cognition’s safety work with customers, policymakers, and the broader research community. ### Exceptional candidates have demonstrated - **Frontier lab safety/alignment experience:** Hands-on safety, alignment, evaluations, or red-teaming at a frontier AI lab; shipped safety work impacting real models or products. - **Agentic systems depth:** Understand how agents fail in practice (tool misuse, specification gaming, long-horizon drift, adversarial inputs) and have ideas to measure and mitigate. - **Research credibility:** Published or widely used safety research, evals, or methods; an advanced degree in CS/ML or related field is a plus. - **Technical fluency:** Comfortable in **Python** and the training/inference stack; can read code, run evals, and discuss details with researchers. - **Builder mindset:** Prefer shipping mitigations over writing memos. - **Judgment under uncertainty:** Can make clear deployment-risk calls with incomplete information and explain them to engineers, executives, and customers. ### Equal opportunity Cognition is an equal opportunity employer. They do not discriminate based on protected characteristics and are committed to providing reasonable accommodations for candidates with disabilities—please let them know if you need any.
Listing freshness
CronJobs last confirmed this listing 2h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.