AI Engineer
Make Harold smarter. Ship the intelligence layer. Work at the frontier.
About the role
Harold runs on AI. Not as a gimmick but as the actual product. Every appraisal, every verdict, every insight Harold surfaces is powered by an AI pipeline that we are building in real time. We need someone who lives and breathes this stuff to help us build it properly.
You will work directly with the CTO on the AI layer that makes Harold genuinely intelligent — the models, the pipelines, the prompts, the evals, the infrastructure that turns raw deal data into the smartest analysis in the room.
What you’ll be doing
Design, build, and improve the AI pipelines that power Harold's core intelligence
Own prompt engineering, model selection, evaluation, and continuous improvement
Build agentic workflows that make Harold proactively useful — not just reactive
Integrate new models, APIs, and tools as the landscape moves — and it moves fast
Work end to end from the model layer to the product surface where users see the output
About you
You are obsessed with what LLMs can actually do and you push them further than most people think possible
You have built AI-powered products or pipelines before — not just called the OpenAI API once
You follow the frontier closely — new models, new techniques, new papers — and you apply it fast
You write clean code but you know that shipping beats perfection every time
You are as comfortable in a system prompt as you are in a codebase
You want to work somewhere your AI work actually matters and reaches real users
Benefits
Equity and token allocation - be part of what Harold becomes
Direct collaboration with the CTO on the most interesting AI problems in property tech
Access to the full model stack - Anthropic, OpenAI, whatever the best tool is
Remote-flexible with a London base
Responsibility and ownership from day one. No junior queues, no PR approval chains
A product that enterprise clients actually depend on
AI
London / Manchester
Remote
Full-time
Competitive
