Staff Software Engineer, AI Reliability

About the role

Anthropic is looking for a Staff Software Engineer to join its AI Reliability Engineering (AIRE) team. You will work across the entire serving stack – from SDKs to network, API layers, serving infrastructure, and accelerators – to make Claude’s delivery more robust and resilient.

Key responsibilities

• Develop Service Level Objectives for large‑language‑model serving systems, balancing availability, latency, and development velocity.
• Design and implement monitoring and observability solutions across the token path.
• Assist in building high‑availability serving infrastructure across multiple regions and cloud providers.
• Lead incident response for critical AI services, ensuring rapid recovery and systematic post‑mortem improvements.
• Support the reliability of safeguard model serving to meet safety commitments.

Required profile

• Strong background in distributed systems, infrastructure, or reliability engineering.
• Curious and brave, able to dive into unfamiliar systems during incidents.
• Holistic view of system composition and seam identification.
• Excellent communication and collaboration skills, capable of building lasting cross‑team relationships.
• Ownership mindset for outcomes, even for systems you do not directly own.

Required skills

• Distributed systems
• Infrastructure engineering
• Reliability engineering / SRE practices
• Monitoring and observability tools
• High‑availability serving architectures
• Cloud platforms (multi‑region deployment)
• GPU, TPU, Trainium accelerators
• ML‑specific networking optimizations such as RDMA