Senior ML Engineer: AI Safety & Alignment (RLHF)

AI Breaking Wire; San Francisco, Northern (CA, KY); Hybrid; USD 250,000 - 400,000; Full time

Benefits offered by this job

Top-tier salary and equity grants

Comprehensive medical, dental, and eye

Job summary

Anthropic, based in San Francisco, is seeking a Senior Machine Learning Engineer to advance AI safety and alignment as our models scale. You will help build reliable, interpretable, and steerable systems that behave safely and beneficially.

You will implement and scale alignment techniques such as RLHF and Constitutional AI, develop automated evaluation pipelines for vulnerabilities, and partner with policy and research teams to embed safety standards into model training.

Qualifications

• 4+ years of industry experience building and deploying ML models at scale.
• Strong proficiency in Python, PyTorch, and distributed computing frameworks (e.g., Ray, DeepSpeed).
• Deep familiarity with safety challenges in generative AI and large language models.

Responsibilities

• Implement and scale alignment techniques such as Reinforcement Learning from Human Feedback (RLHF) and Constitutional AI.
• Develop automated evaluation frameworks to detect vulnerabilities, hallucinations, and harmful outputs.
• Partner with policy and research teams to operationalize safety standards into model training pipelines.
• Optimize training infrastructure for safety-focused objective functions.

Skills

Python

PyTorch

Reinforcement learning

AI safety

Education

Ph.D. in Computer Science

M.S. in Computer Science

B.S. in Computer Science

Tools

Ray

DeepSpeed

Job description

Anthropic, based in San Francisco, is seeking a Senior Machine Learning Engineer to advance AI safety and alignment as our models scale. You will help build reliable, interpretable, and steerable systems that behave safely and beneficially.

You will implement and scale alignment techniques such as RLHF and Constitutional AI, develop automated evaluation pipelines for vulnerabilities, and partner with policy and research teams to embed safety standards into model training.