Junior Research Scientist (AI Evaluation)

Junior Research Scientist (AI Testing and Evaluation) Salary: 45k Location: London, UK Type: Full-Time, Permanent Work Model: Hybrid (Minimum 2 days a week in the office) About Us Advai works with enterprises, government, and defence organisations to test AI systems before they're deployed. AI adoption is accelerating across the economy, government, and defence, and our mission is to enable that — giving organisations the confidence to deploy AI safely and securely through evidence-based testing. We're growing our research team and looking for motivated, curious Junior Research Scientists to work on high-impact AI evaluation across commercial, public sector, and defence use cases. Role Overview This is a hands-on role for someone with a scientific mindset and the communication skills to help customers understand their AI. As a research scientist at Advai, you'll work directly with customers to make sense of their AI systems, then design, build, and run the tests that show whether those systems are safe to deploy. It's a genuinely varied role — a mix of scientific, engineering, and advisory work — with real impact on how AI gets deployed across industry, government, and defence. If you enjoy both deep technical problems and working closely with people, this is a role where you'll see your work make a difference heloig our customers deploy AI that they can trust. What You Will Actually Be Doing Your core work is taking a customer's AI use case and driving it all the way through to tested results on our platform: Onboard use cases: work with customers to understand their AI systems and bring those use cases into Advai's testing product. Understand risk and requirements: help customers articulate their risks and testing requirements, and turn vague concerns into clear, testable statements. Define tests: translate those requirements into concrete test definitions — the data, scorers and procedures needed to evaluate the system. Implement and run tests: instantiate those definitions as procedures on the Advai platform, run them, and produce results. Integrate into the platform: work with our engineering team to build testing procedures into the Advai testing platform. Contribute to results delivery: help shape the front-end design that presents results back to customers clearly. Alongside this core delivery, of research scientists also provide expert advisory: Governance and assurance advice: support customers with governance, risk management, and AI assurance processes. Interpreting results: help customers make sense of test results and advise on next steps and future development. AI expertise: act as a trusted expert for other AI questions as they arise. You can also contribute to internal research projects and help shape the next generation of Advai's automated testing and monitoring capabilities. Requirements - Bachelor's or Master's degree in Machine Learning, Computer Science, or simiarlly reelvent field. - Practical experience with Python and common AI/ML frameworks (e.g., PyTorch, Huggingface, frontier AI APIs). - A genuine passion for AI testing, evaluation, and monitoring, and how AI is being applied across real-world use cases. - Confidence engaging directly with customers and external stakeholders — understanding their requirements, explaining technical concepts, and advising them. - Willingness to work across the full stack of the job: scientific test design, hands-on engineering, and client-facing delivery. - Curiosity and willingness to explore new research areas and learn quickly. - Strong collaborative skills and ability to work effectively across research and engineering teams. - Good written and verbal communication skills for technical reports and client-facing engagements. Nice to Have - Exposure to adversarial AI, red teaming, robustness testing, or AI security evaluation. - Experience with Linux and remote computing environments. - Familiarity with software engineering practices such as version control and testing frameworks. - Experience advising on governance, risk, or assurance processes. Benefits - Flexible working hours with a hybrid model, blending in-office collaboration in London with the flexibility of remote work. - Private health and cycle to work scheme. - Opportunities for professional development in the rapidly evolving field of AI. - A dynamic and exciting work environment focused on enabling other companies to deploy AI with confidence. We are committed to fostering a diverse and inclusive workplace. We strongly encourage qualified candidates from all backgrounds to apply.