AI Penetration Tester

@ Crossing Hurdles

About Crossing Hurdles

At Crossing Hurdles, we work as a referral partner, connecting candidates with a world-leading AI research lab partner that builds and trains cutting-edge AI models.

Position: AI Penetration Tester

This role focuses on red-teaming AI models and agents by crafting jailbreaks, prompt injections, misuse cases, and exploit scenarios. Training support is provided.

Key Responsibilities

• Red-team AI models using adversarial techniques.
• Generate human data by annotating AI failures.
• Apply structured testing using taxonomies and benchmarks.
• Document reproducible reports, datasets, and attack cases.
• Support multiple projects including LLM jailbreaks and socio-technical abuse testing.

Required Qualifications

• Prior red-teaming, cybersecurity, or adversarial AI experience.
• Expertise in adversarial machine learning, including jailbreak datasets and prompt injections.
• Cybersecurity skills such as penetration testing and exploit development.
• Experience with socio-technical risks such as disinformation or abuse analysis.
• Creative probing skills using psychology, acting, or writing.

Application Process

Submit your resume, complete an AI interview based on your resume (15 min), and fill out a brief application form (20 min).

Key skills/competency

• Adversarial AI
• Penetration Testing
• Red-teaming
• Cybersecurity
• Exploit Development
• Reverse Engineering
• Jailbreaks
• Prompt Injection
• Data Annotation
• Risk Analysis