Want to get hired at Crossing Hurdles?

AI Red-Teamer - Adversarial AI Testing

Crossing Hurdles

HybridHybrid

Original Job Summary

About the Role

The AI Red-Teamer - Adversarial AI Testing position at Crossing Hurdles is a contract role where you will red-team AI models and agents through innovative adversarial techniques. This role involves crafting jailbreaks, prompt injections, misuse cases, and exploit scenarios to test and secure cutting-edge AI models across various projects.

Key Responsibilities

  • Red-team AI models via creative adversarial testing
  • Generate high-quality human data by annotating failures
  • Apply structured taxonomies, benchmarks, and playbooks
  • Document findings with reproducible reports and datasets
  • Support multiple projects including LLM jailbreaks and socio-technical abuse testing

Required Qualifications

  • Experience in AI adversarial work, cybersecurity, or socio-technical probing
  • Expertise in adversarial machine learning including jailbreak and prompt injection techniques
  • Cybersecurity skills such as penetration testing and exploit development
  • Background in socio-technical risk analysis and creative probing strategies

Application Process

Complete a 20-minute resume upload, a 15-minute AI interview based on your resume, and submit a form.

Key skills/competency

Adversarial Testing, Cybersecurity, AI, Red-Teaming, Penetration Testing, Exploit Development, Socio-Technical Analysis, Prompt Injection, Jailbreak, Risk Analysis

How to Get Hired at Crossing Hurdles

🎯 Tips for Getting Hired

  • Customize your resume: Tailor it to highlight adversarial AI experience.
  • Align with security skills: Emphasize cybersecurity and red-teaming expertise.
  • Demonstrate practical projects: Showcase real testing and report samples.
  • Research Crossing Hurdles: Understand their partnerships and AI model context.

📝 Interview Preparation Advice

Technical Preparation

Review adversarial machine learning techniques.
Practice penetration and exploit testing methods.
Study AI model vulnerabilities and benchmarks.
Update skills on prompt injection and jailbreak scenarios.

Behavioral Questions

Describe a challenge in red-teaming.
Explain handling unexpected AI failures.
Share experience with time-constrained projects.
Discuss teamwork in remote assignments.