Intern for AI Security RL Gyms

Alice builds adversarial evaluation infrastructure used by the world's leading AI labs to stress-test their most capable agents before deployment. We're hiring interns for our RL Gym environments: adversarial training grounds where AI agents face prompt injection, social engineering, and data exfiltration scenarios.

The Role: You design prompt injection scenarios in YAML, run them against frontier models, validate success rates, and submit passing tasks.

The Task

You build adversarial prompt injection tasks for Alice's RL Gym platform. Each task is a self-contained YAML scenario simulating a realistic AI agent deployment, testing whether the agent can be manipulated into violating its safety policies.

What a task includes:

1. An agent persona with a system prompt, security policies, and a defined set of tools.
2. A simulated environment (e-commerce site, messaging platform, enterprise app) with content the agent interacts with.
3. An adversarial payload embedded in the environment (in messages, product listings, tool responses, documents) that attempts to trick the agent into a policy violation.
4. A benign user request that naturally leads the agent to encounter the payload.
5. Deterministic evaluation criteria specifying exactly what constitutes success (agent resists) vs. failure (agent is compromised).

What We're Looking For

• Strong problem-solving skills and curiosity about AI security.
• Interest in adversarial thinking and understanding how AI agents can be manipulated through prompt injection or other attack techniques.
• Basic understanding of prompt injection concepts (or willingness to learn).
• Comfortable writing structured content in YAML or able to learn quickly.
• Familiar with using the command line (CLI); experience with Docker is a plus but not required.
• Detail-oriented and able to follow technical guidelines consistently.
• Good command of English.
• Background in Computer Science, Cybersecurity, AI, Software Engineering, or related fields is preferred.

What We Offer

• Hands-on experience working on one of the most cutting-edge AI safety and security projects.
• Internship Allowance.
• Mentorship from experienced AI security and red teaming professionals.
• Opportunity to contribute to evaluation environments used by leading AI labs.

If you’re eager to learn, innovate, and grow in the field of data engineering, we’d love to hear from you. Apply today to be part of a team that values creativity and technical excellence!

Please note that only shortlisted candidates will be contacted.

ActiveFence is the leading tool stack for Trust & Safety teams, worldwide. By relying on ActiveFence’s end-to-end solution, Trust & Safety teams – of all sizes – can keep users safe from the widest spectrum of online harms, unwanted content, and malicious behavior, including child safety, disinformation, fraud, hate speech, terror, nudity, and more.

Using cutting-edge AI and a team of world-class subject-matter experts to continuously collect, analyze, and contextualize data, ActiveFence ensures that in an ever-changing world, customers are always two steps ahead of bad actors. As a result, Trust & Safety teams can be proactive and provide maximum protection to users across a multitude of abuse areas in 70+ languages.

Backed by leading Silicon Valley investors such as CRV and Norwest, ActiveFence has raised $100M to date; employs 300 people worldwide, and has contributed to the online safety of billions of users across the globe.