Artificial Intelligence Engineer

London | 1-2 days a month in office

Join a founding team from Ares, Goldman Sachs and Deloitte building AI agents to automate real estate credit underwriting.

Today, highly paid analysts can spend 1–2 weeks working through messy PDFs, spreadsheets, market data and financial models to screen a deal.

This startup is building agents that do that work in around an hour – reading documents, researching markets, reasoning across information and producing analysis where every figure traces back to its source or calculation.

The product is already in production with 3 funds, including 2 on paid 12-month contracts.

They're now hiring their first engineer. You'll take over the codebase from the CTO and help choose the next engineers they hire.

You'll work on:

• Evals against analyst-reviewed ground truth
• Human-in-the-loop agents that can pause + resume
• Agents that build + iterate auditable Excel financial models
• Document pipelines across messy PDFs, charts + spreadsheets
• Python/FastAPI backend + React/TypeScript frontend

The key requirement is that you've already built and run LLM agent systems in production – dealing with retrieval, context, tools and orchestration yourself.

They want someone who brings the war stories: what broke, why it broke and how you made agents reliable enough for real users to depend on.

Strong production Python is essential, alongside enough React/TypeScript to own features end-to-end.

Experience with evals, document extraction, finance or another environment where accuracy and auditability matter is particularly relevant.

£120k–£140k + ~2% equity. The team works together in Soho on Mondays, but for the right person 1–2 days a month is fine.

• Conversation with the CTO
• Architecture/case-study discussion around real problems they're solving – no coding
• Walk through something you've built + product demo + meet the founding team

#J-18808-Ljbffr

Full job description from WhatJobs. Apply for this role →