Senior Software Engineer

Building the AI infrastructure layer for how enterprises use their data to power models, agents, and workflows.💫 Company | AI-native developer platform / AI infrastructure📈 Stage | Series C, $60m ARR, 30k customers, 50% YoY growth📏 Size | ~300 globally, London office recently opened🧢 Role | Senior/Staff Software Engineer🎯 Areas | Agentic systems, LLMs, MCP, evals, prompt engineering✨ Skills | TypeScript / React, full stack product development📍 Based | Hoxton, London💻 Hybrid | 3 days in-office💰 Offer | £120-150k + stock options (pre-IPO)BackgroundMost companies are asking how to use AI. This one already has the answer, and is building it into a product customers pay for.Their platform holds one of the most valuable things you can feed a LLM: structured, governed enterprise data at scale. 30k companies, including some of the biggest names in tech, run on it. While competitors bolt AI onto legacy systems, this was built for the AI era from day one.They're now hiring 2-3 Senior/Staff Software Engineers to join the small team at the centre of it all.The team buildsAn AI assistant that deeply understands each customer's data - relationships, rules, business logic. It can interrogate thousands of records in seconds and generate perfectly structured output.The goal now: turn it from a great feature into a revenue line.The role isFull-stack product development, not narrow ML.React interfaces, APIs, backend services, plus the harder layer underneath: agent orchestration, system prompts (treated here as real engineering, not an afterthought), evals, and the query language that gives the agent its understanding of customer data.Think, real-time data graphs, agentic workflows, multi-tenant architecture, MCP integrations.This is forAn engineer strong in TypeScript and React, with LLM systems shipped in production. You know getting a model to do something once is easy - getting it right for every user is the challengeA systems thinker: given a new agent capability, you map user intent, edge cases, failure modes, and what the agent should know, use, and escalateA genuinely good reader who cares how English is put together - why one phrasing lands and another doesn't, what "you" refers to, what's implied vs stated outright. An agent's instructions are prose read literally by a system, and writing them well takes a real feel for language, not just correct grammar. A background in linguistics / languages / classics is a strong signal hereSomeone comfortable with probabilistic systems: you run experiments, design evals, read results statisticallyReach out or apply below and I'll share more about the company and the role.