Staff+ ML Safety Infrastructure Engineer

Anthropic; San Francisco (CA); On-site; USD 180,000 - 280,000; Full time

Benefits offered by this job

Health insurance

Dental insurance

Vision insurance

Parental leave

Equity

Retirement plan

Wellness stipend

Relocation support

Education stipend

Home office stipend

Meals in office

Job summary

Anthropic is seeking an experienced engineer to design and operate scalable ML infrastructure that powers Claude’s safety systems. You will join the Safeguards ML Inference Path team and collaborate with researchers and inference engineers to productionize safety research, optimize inference latency, and ensure reliability at scale.

You will build monitoring, observability, and automated deployment frameworks, enabling researchers to roll out classifiers and models across platforms while

Qualifications

• Proficiency in Python for ML tooling and production pipelines.
• Experience with ML frameworks (PyTorch, TensorFlow, JAX).
• Strong background in distributed systems and reliability for safety-critical deployments.

Responsibilities

• Design scalable ML infrastructure for real-time safety deployments.
• Build monitoring/observability for classifier performance and data quality.
• Collaborate with research to productionize safety techniques.
• Reduce inference latency while maintaining safety standards.
• Implement automated testing, deployment, and rollback for ML models.
• Work with Safeguards, Security, and Alignment teams on requirements.
• Develop internal tools to accelerate safety research and deployment.

Skills

Python

PyTorch

TensorFlow

JAX

Distributed systems

Reliability

Tools

A/B testing frameworks

Experimentation infra

Job description

Anthropic is seeking an experienced engineer to design and operate scalable ML infrastructure that powers Claude’s safety systems. You will join the Safeguards ML Inference Path team and collaborate with researchers and inference engineers to productionize safety research, optimize inference latency, and ensure reliability at scale.

You will build monitoring, observability, and automated deployment frameworks, enabling researchers to roll out classifiers and models across platforms while