AI Safety Expert for English and Punjabi

This position focuses on enhancing the safety of AI systems through proactive testing and vulnerability assessment. As part of a dedicated red team, you will engage in adversarial testing of AI models, identifying weaknesses and generating crucial data that contributes to the overall safety of AI for our customers.

• Red team conversational AI models and agents, including jailbreaks, prompt injections, misuse cases, bias exploitation, and multi-turn manipulation.
• Generate high-quality human data by annotating failures, classifying vulnerabilities, and flagging systemic risks.
• Apply structured methodologies by following taxonomies, benchmarks, and playbooks to ensure consistent testing.
• Document findings reproducibly, producing reports, datasets, and attack cases that customers can act on.

• Prior experience in red teaming, including AI adversarial work, cybersecurity, or socio-technical probing.
• A curious and adversarial mindset, with a tendency to push systems to their limits.
• Structured approach to testing, utilizing frameworks or benchmarks rather than random hacks.
• Strong communication skills to clearly explain risks to both technical and non-technical stakeholders.
• Adaptability to thrive in a dynamic environment with varying projects and customers.

• Experience with adversarial machine learning, including jailbreak datasets, prompt injection, RLHF/DPO attacks, and model extraction.
• Background in cybersecurity, including penetration testing, exploit development, and reverse engineering.
• Knowledge of socio-technical risks, such as harassment/disinformation probing, abuse analysis, and conversational AI testing.
• Creative probing skills, with interests in psychology, acting, or writing that foster unconventional adversarial thinking.

• Identifying vulnerabilities that automated tests overlook.
• Delivering reproducible artifacts that enhance the security of customer AI systems.
• Expanding evaluation coverage by testing more scenarios and reducing surprises in production.
• Building trust among customers regarding the safety of their AI systems through thorough adversarial probing.

Gain valuable experience in human data-driven AI red teaming at the forefront of safety, and play a direct role in making AI systems more robust, safe, and trustworthy.