Senior Machine Learning Engineer, Alignment and Safety

AI Breaking Wire; San Francisco (CA); On-site; USD 180,000 - 260,000; Full time

Benefits offered by this job

Equity compensation

Competitive salary

Remote and hybrid options

Catered lunches and wellness stipends

Job summary

Anthropic is seeking a Senior Machine Learning Engineer to advance Constitutional AI and model alignment. You will scale alignment pipelines to keep Claude safe, helpful, and honest as capabilities grow.

Join a team partnering with researchers to translate alignment theory into production training loops, develop scalable RLHF systems, and build automated evaluations for biases and vulnerabilities.

The role offers flexible remote/hybrid work and competitive compensation.

Qualifications

• 4+ years of industry experience building, training, and deploying large-scale ML models.
• Strong experience with PyTorch and distributed training infrastructure (Megatron-LM, DeepSpeed, or similar).
• Solid foundation in alignment methods, RLHF, and automated red-teaming.
• Passion for AI safety and responsible development of AGI.

Responsibilities

• Build and optimize scalable RLHF and constitutional AI training loops.
• Develop automated evaluation frameworks to detect model vulnerabilities, bias, and undesirable behaviors.
• Partner with research scientists to translate alignment techniques into production-grade training systems.

Skills

RLHF experience

PyTorch

Distributed training

Alignment methodologies

Tools

Megatron-LM

DeepSpeed

Job description

About the Role

At Anthropic, we build reliable, interpretable, and steerable AI systems. We are looking for a Senior Machine Learning Engineer to focus on Constitutional AI and model alignment. In this role, you will scale our alignment pipelines to ensure our Claude models remain safe, helpful, and honest as they grow more capable.

Responsibilities

• Build and optimize scalable reinforcement learning from human feedback (RLHF) and constitutional AI training loops.
• Develop automated evaluation frameworks to detect model vulnerabilities, bias, and undesirable behaviors.
• Partner with research scientists to translate theoretical alignment techniques into production-grade training systems.

Requirements

• 4+ years of industry experience building, training, and deploying large-scale machine learning models.
• Extensive experience with PyTorch and distributed training infrastructure (Megatron-LM, DeepSpeed, or similar).
• Strong foundation in alignment methodologies, RLHF, and automated red-teaming.
• Passion for AI safety and responsible development of AGI.

Benefits

• Top-tier medical, dental, and vision benefits.
• Generous equity allocation and competitive base salaries.
• Flexible work culture with remote and hybrid options.
• Catered lunches and wellness stipends.