Staff+ Application Security Engineer

Staff-plus engineer on Anthropic's Application Security team, which secures the systems that build and serve Claude, including multi-agent orchestration, sandboxed code execution and agents holding delegated credentials. The team uses Claude as its main tool, and this engineer designs, builds and owns Claude-powered security systems such as LLM-driven code analysis, automated vulnerability fixes and AI-assisted threat modelling. The role also leads secure design reviews for new AI systems, evolves the public bug bounty program, embeds with product, infrastructure and research teams as a security owner, and shares an operational rotation covering escalations, incidents and launch consults.

The team wants hands-on application and infrastructure security in cloud and container environments, production code in Python, Go, Rust or TypeScript, practical threat modelling and bug-finding skills, and comfort with high autonomy. Remote-friendly, with travel.