We're hiring a developer to work on agentic AI for engineering compliance.
We build AI agents that verify designs against complex technical requirements. Think multi-agent systems with sophisticated tool use, explicit reasoning chains, and the kind of auditability needed to satisfy risk-averse design authorities.
This is genuinely hard. Every wall in our office is a whiteboard because we spend a lot of time architecting agentic systems that can act independently, reliably, and safely. Reliable agents are difficult to build, scale, and validate. If you've tried, you know. If you haven't, this isn't the right role — you will be in the deep end from day 1.
What we need:
Someone who is AI-native to join the
Founding Team
. You will already be building agentic systems — skill and state management, chained tool calling, debugging non-deterministic behaviour. Experience with LLM evals and understanding how/why agents fail. Cloud deployment. Modern agentic development workflows (Claude Code, Cursor, spec-driven approaches).
Knowledge graphs useful. Full stack useful. RAG is… a given. Civil engineering background not required — you will get a lot of that here.
What you'll get:
Direct mentorship from technical leadership with deep expertise in AI for civil engineering and a background in academic research. Exposure to business strategy — not just code. Real ownership of hard problems, not tickets. Latest tooling, no compromises. EMI equity — you'll own part of what we're building.
The location:
Wallingford, Oxfordshire. In-office 4-5 days. No hybrid. The problems we are working on are very difficult and you will not absorb the concepts if you are not physically present.
Salary:
£35-50k + EMI equity depending on experience. We're open to strong graduates, career-switchers, or developers early in their career who've already been building with these tools.
The interview process:
- Application
— CV and a link to something you've built. We'll review within a week.
- Algorithm screen
— A short HackerRank assessment. We're not looking for LeetCode grinders, but you need solid fundamentals in graphs, state machines, and complexity reasoning.
- Architecture question
— A written take-home asking you to design an agentic system. No code, just thinking. ~60 minutes.
- On-site technical
— Whiteboard architecture, then build it live. AI tools allowed — we want to see how you actually work. Lateral thinking required. If you cannot move quickly and safely under pressure, please do not apply.
We're a small team and client pull means we need to grow. If you want to work on state of the art systems with with a very strong technical team, send your CV and a link to something you've built to info@inframatic.ai
No recruiters.
