Staff Plus Software Engineer, Anthropic

• Proficiency in Python and SQL
• Experience building and operating data pipelines or data stores in production
• Ability to work across the data stack, including ingestion, storage, and consumption
• Strong written and verbal communication skills, including the ability to explain technical tradeoffs to people outside your discipline
• We encourage you to apply even if you do not believe you meet every single qualification. Not all strong candidates will meet every single qualification as listed

Desirable

• We are looking for software engineers to help build safety and oversight mechanisms for our AI systems. As a software engineer on the Safeguards team, you will work to monitor models, prevent misuse, and support user well-being
• This role focuses on the data foundations that make that work possible: the pipelines, stores, and governance controls that our detection, evaluation, and review systems are built on, deployed across AWS, GCP, and Azure
• You will apply your technical skills to uphold our principles of safety, transparency, and oversight while enforcing our terms of service and acceptable use policies
• The systems you build sit underneath decisions with real consequences for customers and for people affected by model misuse, and much of the data moving through them is sensitive. Correctness, retention discipline, and access control are core engineering requirements here rather than things bolted on afterward. You will also design for portability from the outset, because these systems run in customer-managed and third-party environments you do not fully control
• Build and operate the data platform that powers Safeguards, including ingestion and processing pipelines, warehouses and other data stores, and the schemas and interfaces that detection and review systems depend on
• Keep Safeguards systems running day to day and hold a high operational bar that serves both safety and customers, while reducing the manual effort needed to sustain it
• Own data governance and integrity, including retention and access controls, privacy-preserving handling of sensitive data, lineage, and correctness guarantees that downstream consumers can rely on
• Design systems that run portably across cloud providers, working within the constraints of customer-managed and third-party environments
• Partner with the analysts, investigators, and researchers who rely on this data, and build the internal tooling their work depends on