Reinforcement Learning Engineer

What you'll do

• Build RLVR environments, verification pipelines, and deterministic reward signals from real-world vulnerabilities.
• Create secure, reproducible sandboxes and scalable infrastructure for agent-driven exploitation, patching, and distributed training.
• Benchmark AI models on cybersecurity tasks and collaborate with frontier AI labs on evaluation formats and metrics.

What they ask for

• Understanding of RL training workflows for modern LLM systems, especially execution-based feedback or RLVR.
• Proficiency in Python and C.
• Solid understanding of software vulnerabilities, binary exploitation, fuzzing, or program analysis.
• Experience with DevOps pipelines, reproducible builds, Linux, low-level debugging, or benchmark environments.

Nice to have

• Rust experience.
• Experience with reward functions or verifiers, low-level program analysis tools, or cybersecurity benchmarks and CTFs.

Pay

$176,400 - $242,550 USD base salary

Tools

Python, C, Rust, GitHub Actions, Docker, BuildKit, Nix, Linux

AI focus

Develops verifiable reinforcement-learning environments and benchmarks for AI agents learning cybersecurity tasks, including vulnerability discovery, exploitation, and remediation.