Security Labs Engineer

About this role

Anthropic's Security Labs team tackles high-risk, high-impact AI security research through short-cycle experimental projects. You'll lead ambitious bets on frontier security challenges—from cryptographic verification to hardened infrastructure—with the understanding that meaningful failure is part of the process.

What you'll do

• Own security research projects end-to-end, from scoping to prototype to hand-off or documented closure
• Build novel security infrastructure rapidly (isolated clusters, attestation systems, hypervisor solutions) optimized for learning
• Collaborate with research and infrastructure teams to validate whether proposed security measures work with real workloads
• Select and prioritize future project bets based on technical merit and strategic impact
• Translate experimental results into architectural recommendations and contingency plans for AI security
• Document findings and create writeups that inform long-term security strategy

What they're looking for

• Strong software engineering fundamentals
• Deep expertise in at least one specialized area: firmware/hardware security, cryptography, kernel/hypervisor internals, formal methods, or reverse engineering
• Systems-level thinking and ability to jump between architectural layers
• Self-directed project ownership and prototyping
• Comfort with ambiguity and high-risk technical bets
• Cross-team collaboration and communication
• Applied security research mindset