Bugcrowd
Reinforcement Learning Engineer (Cybersecurity)
About this role
Build infrastructure and tooling to create reinforcement learning environments that train AI systems in cybersecurity tasks like vulnerability discovery and exploitation. This systems engineering role focuses on designing pipelines that transform real-world vulnerability research into large-scale training environments for frontier AI labs.
What you'll do
- Design and build pipelines that ingest software projects and automatically construct RL training environments
- Integrate Bugcrowd's Mayhem platform with environment generation workflows
- Develop reproducible, containerized Linux ML environments for RL training
- Create systems that scale to thousands of training scenarios for AI model development
- Work with binary exploitation and fuzzing to inform environment design
- Collaborate with frontier AI labs including Anthropic, OpenAI, and Cohere on environment specifications
What they're looking for
- Reinforcement learning workflows and LLM training pipelines
- Python and C development; Rust a plus
- Linux systems administration and low-level debugging
- Docker, containerization, and reproducible build systems (Nix, Buildkit)
- Binary exploitation, fuzzing, and vulnerability analysis
- DevOps pipelines and CI/CD (GitHub Actions)
- Experience with benchmark environments (CTFs, SWE-bench, security challenges)
- Large open-source codebase familiarity
Opens the official application on the employer’s site. No login required.
Bugcrowd
Bugcrowd builds infrastructure that converts vulnerability research into AI training environments and develops advanced exploit capabilities for cybersecurity applications. The company is hiring for senior engineering roles including staff engineers to design AI training pipelines and cleared vulnerability researchers to conduct exploit development and autonomous security research.
- Website
- bugcrowd.com
Likely interview questions
- Describe your experience building reproducible machine learning environments and how you've handled environment consistency across different systems.
- Walk us through a project where you integrated vulnerability analysis tools into an automated pipeline—what were the main challenges?