Senior AI Engineer

About LEC AILEC AI is the AI business inside LEC Industries, a London-headquartered group building technology and operating companies across robotics, AI and consumer brands, with seven decades of operating depth between the UK and China. We are not a research lab and not a corporate IT function. We are a tech business building tech, with live commercial deployments and real users from day one.We build two core products: Donnie, our agentic operating layer for organisations (persistent memory, multi-agent orchestration, tool use, digital twins), and Bishop, our proprietary data foundation model for structured and tabular data (forecasting, prediction, anomaly detection). On top of these we build applied products and run the AI transformation of companies.The RoleWe are hiring a Senior AI Engineer to work directly with our team and our CEO, and to own products end to end across everything we build: Donnie, Bishop, our marketing and content platform, and the applied products we take to market.This is a senior, deeply hands-on builder-architect seat. You design the system and you ship it. You take a problem from a blank page to a deployed product in weeks and own it in production.This is a founding-team-level opportunity. You operate with the autonomy, ownership and pace of a founder, shaping the products, the stack and the company itself as we grow from a handful of engineers to the team that takes this to market. You will sit at the centre of what we build, with direct access to how a Group CEO actually thinks and sells, and real influence over where this goes.What you will own- Architecture across memory, retrieval, knowledge graphs and agent orchestration, and the reliability of the whole AI stack in production- The data layer end to end: ingestion, cleaning, migration and robust pipelines over messy real-world organisational data- Model work: selection, fine-tuning, retraining and post-training on our own GPUs, moving toward company-specific model training as the data allows- Evaluation pipelines for agentic AI, plus observability, cost controls and latency budgets- Multi-tenant, secure, production-grade systems and the path from internal platform to SaaS- Deployment across our own servers and cloud, and the data governance around it- Turning stakeholder requirements into shipped features fast, and helping take products to marketWho you areYou have shipped agentic AI and LLM systems into production, with real users, real data and real failure modes. Not a demo, not a hackathon. You have strong hands-on depth in most of:- LLM memory and retrieval at real scale (RAG, vector and graph databases)- Real multi-agent orchestration, tool use and MCP, not a chain of LLM calls- Fine-tuning open-weight models on proprietary data and shipping the result- Self-hosting: GPU inference, containerised services, running open-source models locally- MLOps: evaluation, observability, cost tracking, and multi-tenant secure deploymentYou can architect a system and still write code every day. Python is your first language, and you are comfortable across the full stack of getting things live. You read a paper in the morning and ship a prototype by evening. You have strong opinions on evaluation that go beyond vibes, and you favour maintainable over clever.Above all, you are hungry and self-driven. You build because you cannot help it. You take ownership before anyone hands it to you, you move faster than people expect, and you want to be early, own something real, and see it live in the market. You operate without a quarterly roadmap, ship in days, take the brief and run, and lift the engineers around you as you go.Tech stack (high level)Python across the board; modern LLMs with agent and tool-use frameworks; retrieval-augmented generation over vector and graph data stores; containerised, self-hosted infrastructure with GPU inference; and standard MLOps and observability tooling. We care more about strong fundamentals and the ability to learn fast than about any specific tool.Bonus signals- Tabular or foundation-model work relevant to Bishop- AI marketing or content systems- Open-source work other engineers actually use- Audio or vision pipelines shipped at production quality- Mandarin (we operate across the UK and China)Logistics- Full-time, office-based, five days a week. Not remote or hybrid.- Fulham, London.- Right to work in the UK required; we do not provide visa sponsorship.- Compensation: competitive, based on experience.How to applyEmail talent@lecai.ai with your CV, your GitHub or portfolio, and a short note on the most impressive AI system you have taken to production, what you owned, and one decision you would make differently now.