AI Red Team Engineer

This is a fully remote position, open to applicants in France.

📋 Description

• Conduct red-team assessments on LLM-powered systems, including chatbots, copilots, RAG pipelines, AI agents, tool-calling workflows, and API-based AI products.

• Evaluate for vulnerabilities such as jailbreaks, prompt injection, system-prompt and tool leakage, sensitive data and context leakage, unsafe outputs, policy bypass, tool misuse, excessive agency, resource and token-cost exploitation, and business-logic violations.

• Develop lightweight Python scripts to automate attacks, execute prompt sets, access model APIs, gather and evaluate responses, and create repeatable reports.

• Create and sustain an internal attack library containing prompts, scenarios, test cases, regression tests, scoring rubrics, and reusable demo cases.

• Convert model failures into comprehensive reports detailing what occurred, its significance, reproduction steps, severity, and potential fixes.

• Transform successful attacks into regression tests and product specifications.

• Stay updated on emerging red-team and safety techniques, integrating valuable insights into our testing protocols.

• Aid go-to-market efforts by generating strong, credible evidence for customer demonstrations, security audits, and sales discussions.

⛳️ Requirements

• Possess a genuine passion for breaking systems and engaging in adversarial reasoning.

• Have a background in QA automation, Application Security, API/security/penetration testing, or bug bounty programs.

• Exhibit strong Python scripting abilities.

• Have experience in testing APIs, web applications, backends, or SaaS platforms.

• Be hands-on with LLMs, prompts, system instructions, RAG, agents, and tool/function invocation.

• Understand LLM-specific abuse vectors, including prompt injection, jailbreaks, data leakage, tool misuse, excessive agency, and token-cost exhaustion.

• Demonstrate the ability to identify bypasses, exploit edge cases, connect failures, and assess real-world implications.

• Distinguish between genuine customer risks and low-impact prompt tricks.

• Produce clear, reproducible bug reports in straightforward English.

• Be capable of working swiftly without having perfect requirements.

• Maintain a strong ethical stance: engage in red-teaming to enhance system safety, operate within defined scopes and legal boundaries, and avoid producing or distributing genuinely harmful materials.

🏝️ Benefits

• Paid time off in accordance with your local regulations, regardless of your work location.

• Work from Paris (hybrid) with a relocation package available.

• Access to the best medical insurance in France.

• All necessary hardware, tools, and services provided.

• Subscriptions for AI agents covered.

• Team off-sites twice annually: recent locations include the Alps and Saint-Tropez.