Founding AI Engineer | Agentic AI, LLMs, Evals | Production Code Generation, Agent Systems, Model Routing - Tackling a $150BN+ Industry | London, 4/5 days (On-Site) | Up to £150k + Equity

Founding AI Engineer | Agentic AI, LLMs, Evals | Production Code Generation, Agent Systems, Model Routing - Tackling a $150BN+ Industry | London, 4/5 days (On-Site) | Up to £150k + EquityThe CompanyWe're partnering with a well-funded, early-stage AI start-up tackling a huge global industry with an ambitious new product.They're building agentic systems that turn ideas into complex, production-ready software – not another LLM wrapper.The technology is already working. The challenge now is making it more reliable, intelligent and scalable, solving difficult problems across agents, evals, context, performance and inference economics.They're looking for a Founding AI Engineer to own it.The Role | Founding AI Engineer | Agentic AI, LLMs, Evals | Production Code Generation, Agent Systems, Model Routing - Tackling a $150BN+ Industry | London, 4/5 days (On-Site) | Up to £150k + EquityYou'll own the agent system responsible for turning high-level requirements into working, production-quality software.Working directly with the founders and engineering team, you'll design agent loops, build production evals, manage context, route between models and optimise for quality, latency and cost.The challenge isn't getting an LLM to generate code. It's building an autonomous system that generates software that actually works.What You'll Be Doing | Founding AI Engineer | Agentic AI, LLMs, Evals | Production Code Generation, Agent Systems, Model Routing - Tackling a $150BN+ Industry | London, 4/5 days (On-Site) | Up to £150k + EquityOwn the agent pipeline from specification to productionBuild workflows across planning, code generation, review, testing and releaseBuild evals and regression tooling to measure output quality and catch failuresDevelop model routing based on quality, latency and costBuild retrieval and context systemsImprove or replace third-party agent infrastructure where neededExperiment with new models, fine-tuning and post-trainingYou'll own one of the hardest and most important technical problems in the business.Tech & AI Stack | Founding AI Engineer | Agentic AI, LLMs, Evals | Production Code Generation, Agent Systems, Model Routing - Tackling a $150BN+ Industry | London, 4/5 days (On-Site) | Up to £150k + EquityTypeScript | LLMs | Tool-Using Agents | Evals | Model Routing | Retrieval & Context | Fine-Tuning | Code GenerationDeep TypeScript experience isn't essential – strong systems engineering experience in another language is absolutely fine.What They're Looking ForProduction experience building LLM/agentic systemsMulti-step pipelines, tool-using agents or autonomous systemsAgents that produce output that actually has to run and work, not just generate textStrong experience with evals, benchmarks and regression testingStrong systems engineering fundamentalsUnderstanding of the trade-offs between model quality, latency and costSomeone who follows the AI field closely and actively experiments with new models and techniquesNice to have: model routing, fine-tuning/post-training, retrieval systems, proprietary agent runtimes, code-generation agents or developer tooling.This is particularly suited to engineers who have gone beyond chatbots, RAG and simple LLM integrations and built agentic systems that need to perform reliably in production.Interview FocusExpect to go deep on three areas:Production Agents – What have you built where the agent's output actually had to run, and how did you know it worked?Evals – What's in your eval suite, and what has it stopped from shipping?Performance & Cost – What was the slowest or most expensive part of your agent loop, and how did you improve it?Why Join | Founding AI Engineer | Agentic AI, LLMs, Evals | Production Code Generation, Agent Systems, Model Routing - Tackling a $150BN+ Industry | London, 4/5 days (On-Site) | Up to £150k + EquityFounding-level ownership over a core AI systemSolve genuinely difficult problems across agents, evals and model infrastructureBuild proprietary AI infrastructure rather than another LLM wrapperWork directly with the founders and influence technical directionUp to £150k + meaningful equityCentral London, 4–5 days on-siteCandidates must already be UK-based with the right to work in the UK. Sponsorship and relocation aren't available.If you've built agentic systems where the output actually has to work – and want to tackle problems at the edge of what's currently possible with LLMs – we'd love to hear from you.Apply with your CV and we'll be in touch if there's alignment.