Lead Safety Architect for Autonomous AI Agents

Cognition AI; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Cognition AI in San Francisco is seeking a senior Safety/Alignment lead to build the safety function for Devin. You will shape the research agenda, evaluation, red-teaming, and deployment policies, working closely with researchers and engineers.

This hands-on role requires frontier-lab safety experience, deep understanding of agentic systems, and a track record shipping mitigations. You will hire and lead safety researchers, represent Cognition AI with customers and the broader research

Qualifications

• Frontier lab safety or alignment experience with real-model impact.
• Agentic systems depth: understanding failure modes in deployed agents.
• Research credibility: published or widely used safety research or evals.
• Technical fluency: comfortable with Python and ML stacks.
• Builder, not reviewer: shipping mitigations over memos.
• Judgment under uncertainty: clear deployment risk calls.

Responsibilities

• Own safety end to end: set Devin safety strategy and outcomes.
• Build the evaluation program: design evals and red-teaming for risk.
• Shape training and the agent harness: integrate safety into training and usage.
• Set deployment policy: define allowed actions and governance with enterprise.
• Build the team and the external voice: hire leads and represent Cognition AI externally.

Skills

Frontier lab safety experience

Agentic systems depth

Research credibility

Python proficiency

Builder mindset

Judgment under uncertainty

Education

Advanced degree in Computer Science / ML or related field

Job description

Cognition AI in San Francisco is seeking a senior Safety/Alignment lead to build the safety function for Devin. You will shape the research agenda, evaluation, red-teaming, and deployment policies, working closely with researchers and engineers.

This hands-on role requires frontier-lab safety experience, deep understanding of agentic systems, and a track record shipping mitigations. You will hire and lead safety researchers, represent Cognition AI with customers and the broader research