AI Red-Teamer

@ Taskify

About the Role

As an AI Red-Teamer at Taskify, you will proactively identify vulnerabilities in AI systems by using adversarial inputs, creative probing, and structured evaluation techniques. This fully remote role in US-friendly time zones offers high-impact projects exploring the frontiers of AI security and ethical safeguards.

Key Responsibilities

• Conduct adversarial testing on AI models including jailbreaks, prompt injections, and misuse cases.
• Generate high-quality human data by annotating failures and classifying vulnerabilities.
• Apply structured frameworks, taxonomies, benchmarks, and playbooks for consistent testing.
• Produce reproducible reports, datasets, and attack cases for actionable insights.
• Support multiple projects and customers including LLM jailbreaks and socio-technical abuse testing.

Qualifications / Skills Required

• Experience in AI red-teaming, cybersecurity, or adversarial testing.
• Strong analytical and adversarial mindset with instinctive probing capabilities.
• Structured approach to testing using established frameworks and benchmarks.
• Excellent communication skills to articulate risks to varied stakeholders.
• Adaptable to managing multiple projects and domains.

Preferred Qualifications

• Expertise in adversarial ML including jailbreak datasets and prompt injection.
• Experience in penetration testing, exploit development, and reverse engineering.
• Knowledge in socio-technical risk analysis and abuse evaluation methods.
• Creative probing skills influenced by psychology, acting, or unconventional writing.

Work Location & Engagement

This is a remote role with work engagements available full-time or part-time, focused on US-friendly time zones (US, UK, Canada). Participation in high-sensitivity projects is optional and supported with guidelines and wellness resources.

Key skills/competency

• Adversarial Testing
• Cybersecurity
• AI Systems
• Risk Analysis
• Frameworks
• Data Annotation
• Penetration Testing
• Creative Probing
• Reverse Engineering
• Communication