Artificial Intelligence Engineer

Requirements

• This role will suit someone smart, self-starting, technically curious and collaborative who wants their research and engineering work to directly improve how people learn, communicate and care for others.
• An AI engineer who has owned work end to end, from an open research question through experimentation to something running reliably in production.
• You think like a researcher but ship like an engineer, you write clean, well-tested Python, and you build services and APIs (for example with FastAPI) that others can depend on.
• A track record of designing and operating dependable AI workflows, including the pipelines, evaluation harnesses, and monitoring that keep complex multi-step systems running predictably and make them straightforward to debug when they do not.
• Testing and deployment are part of how you work, not an afterthought.
• Real depth with large language models and modern AI pipelines, including prompt design, rigorous evaluation, and fine-tuning, paired with the engineering judgement to know when a model is genuinely ready to ship.
• Hands-on experience in at least one signal-rich domain such as speech and audio or computer vision and motion, with the curiosity and ability to grow across modalities rather than staying in a single lane.
• Someone who is comfortable working through technical uncertainty. The problems are well defined, but the answers are not yet known, so you can frame the right questions, design experiments to resolve them, and make steady progress toward an answer.
• Strong communication skills and the ability to make technical work legible to people who are not engineers, including leadership.
• (Desirable) Experience putting machine learning into real-time or interactive systems, particularly anything involving a 3D or games engine, VR, or live sensor and tracking data.
• (Desirable) Background in multilingual or cross-cultural language work, expressive speech, or building AI that behaves believably for people across different cultures.
• (Desirable) Open-source contributions, publications, or other signals that you operate near the front edge of the field, and a thoughtful view on responsible and inclusive AI.

What the job involves

• We’re now looking for an AI Engineer to lead an initial two-year R&D project focused on building the next generation of emotionally responsive virtual humans for soft-skills training - avatars that can talk, move, listen and react to learners in increasingly realistic, human ways.
• You’ll be joining a positive, ambitious and supportive team working at the intersection of AI, immersive technology, behavioural science and real-world social impact, with the autonomy and ownership to shape a major part of our future product direction.
• Set the technical direction for a multi-year research effort and take it from open questions to shipped capability. You will decide what to build, what to buy, and what to prove out first, then make it real.
• Work across the full modern AI stack, combining large language models with speech and vision so that virtual humans can perceive how a learner is communicating and respond in a way that feels natural and appropriate.
• Build the engineering foundations the rest of the team relies on, including the backend services, APIs, and evaluation systems that let complex AI behave reliably and be measured, debugged, and improved with confidence.
• Turn research into a product. You will prototype quickly, evaluate honestly against clear metrics, and harden what works into something that runs in real time inside a live, interactive experience.
• Bring rigour to data. You will shape how data is gathered, annotated, and validated, with real care for diversity, fairness, and how well models hold up in the messy conditions of real use.
• Make sound build versus buy calls on the models and platforms underpinning the product, balancing quality, cost, and the commercial risk of depending on third parties.
• Collaborate widely across product, engineering, learning design, and behavioural science, and use the autonomy you are given to shape a meaningful part of where the product goes next.