The opportunity
Anthropic is working on frontier AI systems that handle sensitive information at enormous scale. How we protect that data, and how we build privacy into our systems rather than bolting it on afterward, is central to our mission of building AI that is safe and beneficial.
What you'll do
Design and implement privacy-preserving architectures for AI training and: inference systems operating at very large scale, using techniques e.g. differential privacy, federated learning, and secure multi-party computation
Partner with researchers to implement privacy-preserving training methods: that protect user data while maintaining model quality
Build foundational privacy infrastructure, including automated data: discovery, classification, access controls, audit logging, and lifecycle management
Translate regulatory requirements (e.g., GDPR, CCPA, HIPAA, the EU AI Act): into technical implementations and automated compliance controls
Architect data governance systems for tracking data lineage, purpose: limitation, and retention across distributed AI systems
Lead privacy reviews and threat modeling for new models and features,: identifying risks and designing scalable mitigations
What they're looking for
- Hands-on experience with privacy-enhancing technologies (e.g., differential: privacy, homomorphic encryption, secure enclaves, secure multi-party computation)
- Experience building privacy infrastructure or controls for machine learning or AI systems
- Experience establishing a privacy engineering practice, or being an early hire in a function
- Experience with distributed systems and cloud infrastructure at scale
- Experience serving as a technical lead on complex, multi-quarter projects
- Contributions to open-source privacy tooling, privacy research, or industry standards
- + years of experience in a software engineering role, including building and: operating large-scale infrastructure
- + years of experience leading large, complex projects as a technical lead
