AI Adversarial Specialist – Remote (English & Urdu)

About the role

We are looking for an AI Adversarial Specialist to join our remote team and help secure conversational AI models. The role focuses on red‑team testing, data annotation, and delivering actionable insights to improve model safety.

Key responsibilities

• Red‑team conversational AI models and agents to uncover jailbreaks, prompt injections, misuse cases, and bias exploitation.
• Generate high‑quality human data by annotating failures, classifying vulnerabilities, and flagging systemic risks.
• Follow established taxonomies, benchmarks, and playbooks to ensure consistent testing methodology.
• Document findings reproducibly, producing reports, datasets, and attack cases for customer action.
• Work independently and asynchronously to meet deadlines while enhancing AI model performance.

Required profile

• Native fluency in English and Urdu.
• Prior red‑team experience in AI adversarial work, cybersecurity, or socio‑technical probing.
• Strong communication skills to explain risks to both technical and non‑technical stakeholders.
• Ability to manage multiple projects and customers effectively.

Required skills

• Adversarial machine learning techniques (jailbreak datasets, prompt injection, RLHF/DPO attacks, model extraction).
• Penetration testing, exploit development, and reverse engineering.
• Experience with socio‑technical risk assessment (harassment/disinformation probing, abuse analysis).

What we offer

• Contract position with hourly compensation of $20–$22.
• Fully remote work environment.
• Opportunity to contribute to cutting‑edge AI safety research.