Become an AI Red Team Expert at LILT
Join the AI Revolution: Help Us Secure the Future of Translation
At LILT, we're not just building AI; we're building the future of global communication. We use cutting-edge AI, machine translation, and a human-in-the-loop approach to break down language barriers and make information accessible to everyone. We're trusted by industry leaders like Intel, Canva, and the U.S. Department of Defense. If you're passionate about AI security and want to make a real-world impact, we want you on our team.
The Mission:
We're seeking talented and driven freelance AI Red Team experts to help us rigorously test and secure our AI systems. As a key member of our red team, you will be focused on adversarial testing of AI systems – LLMs, multimodal models, inference services, RAG/embeddings, and product integrations. Your mission is to think like an attacker, identify vulnerabilities, and help us build more robust and secure AI solutions.
What You'll Do:
Craft creative and sophisticated prompts and scenarios to test model guardrails.
Explore innovative methods to bypass restrictions and uncover hidden vulnerabilities.
Systematically document findings and outcomes to drive continuous improvement.
Collaborate closely with engineers and safety researchers to share insights and strengthen system defenses.
What We're Looking For:
Key Technical Skills:
Generative AI Expertise: Deep understanding of generative AI models, including LLMs, their architectures, training processes (including prompt engineering, fine-tuning, and RLHF), and potential failure modes.
Cybersecurity & Threat Modeling: Solid experience in cybersecurity principles, threat modeling, vulnerability assessment, and penetration testing. Ability to identify attack vectors and simulate real-world threats.
Data Analysis & NLP: Strong analytical skills to dissect model outputs, identify biases or factual errors, and recognize patterns. A background in Natural Language Processing (NLP) is highly beneficial.
Ethical Hacking Mindset: A commitment to using your skills ethically for defensive purposes and responsible disclosure.
Core