Cybersecurity RL Engineer — Build Safe Defensive AI

Anthropic; York and North Yorkshire; On-site; GBP 90,000 - 130,000; Full time

Benefits offered by this job

Health insurance

Dental insurance

Vision insurance

Parental leave

Flexible paid time off

Mental health support

Equity participation

Relocation support

Home office stipend

Commuter benefits

Education stipend

Job summary

Anthropic seeks a Research Engineer to advance safe AI through cybersecurity-focused RL, defense, and safe model training. You will design RL environments, run experiments, and push work into production training loops while collaborating with researchers, engineers, and cybersecurity specialists across and outside Anthropic.

The role blends research and engineering, requiring domain expertise in cybersecurity and interest in training safe AI models. Remote-friendly with relocation support.

Qualifications

• Strong software engineering skills with a focus on safe AI systems.
• Domain expertise in cybersecurity and defensive security.
• Experience with RL environments, experiments, and evaluations.

Responsibilities

• Design and implement RL environments for safe AI training.
• Conduct experiments, evaluations, and deploy production training runs.
• Collaborate with researchers, engineers, and cybersecurity specialists across and outside Anthropic.

Skills

Cybersecurity research

Machine learning / RL

Software engineering

Fuzzing

Detection & response

RL techniques familiarity

Tools

Python

TensorFlow/PyTorch

Job description

Anthropic seeks a Research Engineer to advance safe AI through cybersecurity-focused RL, defense, and safe model training. You will design RL environments, run experiments, and push work into production training loops while collaborating with researchers, engineers, and cybersecurity specialists across and outside Anthropic.

The role blends research and engineering, requiring domain expertise in cybersecurity and interest in training safe AI models. Remote-friendly with relocation support.