Az állásról
The role focuses on designing privacy-preserving architectures, infrastructure, and governance systems for AI training, inference, analytics, and data protection at scale. The engineer will help embed privacy controls into AI products and enable privacy-preserving development across engineering teams.
Responsibilities
- Design and implement privacy-preserving architectures for AI training and inference systems handling billions of conversations, leveraging differential privacy, federated learning, and secure multi-party computation
- Partner with AI researchers to implement privacy-preserving training methodologies that maintain model quality while protecting user data
- Build foundational privacy infrastructure including automated data discovery, classification, access controls, audit logging, and lifecycle management systems
- Translate complex regulatory requirements into actionable technical implementations and automated compliance controls
- Architect comprehensive data governance platforms for tracking data lineage, purpose limitation, and retention across distributed AI systems
- Lead technical privacy reviews and threat modeling for new AI models and features, identifying risks and architecting scalable mitigations
- Collaborate with product and infrastructure teams to embed privacy controls into Claude's inference systems, user interfaces, and data pipelines
- Develop privacy engineering toolkits and frameworks that enable all engineers to build privacy-preserving features by default
- Design and implement privacy-preserving analytics and measurement systems that provide insights while protecting individual user privacy
- Research and evaluate emerging privacy technologies from academia and industry, contributing to open-source tools and AI privacy standards
- Act as consultant and advocate for privacy best practices as central to the mission of AI safety
Requirements
- 8+ years of professional software engineering experience, not including internships and co-ops
- 5+ years of experience focused on privacy, security, or data protection
- Deep expertise in privacy engineering principles: privacy by design, data minimization, purpose limitation
- Strong programming skills in Python, Go, or similar languages with experience building production systems at scale
- Experience with privacy-enhancing technologies, including differential privacy, homomorphic encryption, and secure enclaves
Conditions
- Location: San Francisco
- Salary: $320,000 - $320,000
How to apply
Apply.
Company and product
Employer: Anthropic.
The role includes collaboration on Claude's inference systems and supports the mission of AI safety.