Role Overview
We are hiring for one of our clients, seeking a LLM Red-Teamer to work on a contract basis.
The role involves identifying vulnerabilities and improving the robustness of large language models through adversarial testing.
Candidates will design and execute red-teaming scenarios to probe model behavior under edge cases and malicious inputs.
Key Responsibilities
- Design and implement red-teaming strategies to uncover weaknesses in LLM outputs, including bias, toxicity, and factual inaccuracies.
- Develop automated and manual testing frameworks to simulate adversarial attacks on AI models.
- Document findings in structured reports, highlighting risks and recommending mitigations.
- Collaborate with AI safety researchers to refine testing protocols based on emerging threats.
- Stay current with advancements in AI security, prompt injection techniques, and jailbreak methodologies.
Required Skills & Qualifications
- Experience with adversarial testing of LLMs, including prompt engineering and attack simulations.
- Proficiency in Python and familiarity with AI/ML frameworks such as Hugging Face or PyTorch.
- Knowledge of cybersecurity principles, particularly those relevant to AI systems.
- Ability to analyze model outputs for bias, toxicity, and hallucinations.
- Strong written communication skills for documenting vulnerabilities and recommendations.
- Familiarity with ethical AI guidelines and responsible AI development practices.
More About the Opportunity
This role offers a unique opportunity to work with a global leader in the Technology, Information and Internet industry, contributing to the advancement of safer AI systems.
The position involves direct impact on model deployment decisions by identifying critical failure modes before public release.
Equal Opportunity Employer
We hire based on skills and expertise.
All qualified candidates are welcome regardless of background, experience, or prior employment history.
Applications are reviewed solely on demonstrated technical ability and qualifications.