Principal AI Research Scientist

The OpportunityOur client, a well-funded AI security company working with leading model developers and enterprises, is hiring a Principal Research Scientist to lead hands-on work on how frontier models are tested, measured and stress-tested. This is a senior individual-contributor role for a researcher who is equally at home writing code as writing papers, and who thrives on tight research loops: framing a question, building the tooling, running the experiments and landing conclusions that customers can act on. You’ll bring deep knowledge of large language models, reasoning systems and agents, and use it to turn messy, real-world customer problems into rigorous, measurable research.What You’ll DoLead hands-on research into frontier models, including reasoning, agentic and multimodal systems.Build and continuously refine benchmarks, evaluation suites and experimental infrastructure.Run fast, focused research sprints that move from hypothesis to prototype to validated result.Surface where models fall short, where current evaluation methods miss the mark, and where new research directions are opening up.Partner directly with customers and their technical teams to frame open-ended problems as concrete research questions.Work alongside engineering and product to turn findings into capabilities that ship.What You’ll BringA PhD or MSc in Computer Science, Machine Learning, AI or a related discipline, backed by substantial research experience.First-author or significant contributions at top venues such as NeurIPS, ICML, ICLR, ACL or comparable.Outstanding engineering ability, including advanced Python and PyTorch, and a track record of building and evaluating sophisticated AI systems.Comfort owning experiments end to end and delivering results on short timelines.Deep expertise in LLMs, foundation models and evaluation methodology.Strong analytical and communication skills, and the ability to stay effective in a fast-moving environment shaped by customer needs.Time at a frontier AI lab is a significant plus.Bonus PointsApplying AI to security problems, for example autonomous security agents, vulnerability discovery, automated penetration testing or SOC workflows.Securing AI systems, including LLM red teaming, adversarial ML, prompt injection and the security of agentic systems.Familiarity with security research environments such as capture-the-flag challenges, sandboxed exploitation labs, static and dynamic code analysis, and security testing frameworks.Building security-focused benchmarks or evaluating frontier models on demanding security tasks.