Description:
We're looking for a graduate data scientist to help build and exploit a digital twin of every block of flats in the UK — a single, structured, continuously updated model of the buildings we insure and the ones we might.
You'll do two things. First, help build the data foundation: bringing together public and commercial datasets and resolving them to individual buildings. Second — and increasingly over time — use that data to understand the true risk of each building and turn it into pricing, risk selection and insight we give back to clients.
Initially, this is a three month role with the potential to become full time.
This is a rare chance to own foundational work from day one at an early-stage company, with real scope to shape both the data and the models built on it.
What you'll do
- Help build and maintain the digital twin — ingesting, cleaning, matching and enriching building-level data from multiple sources.
- Work with UK datasets including Ordnance Survey (e.g. AddressBase, building footprints), Land Registry, EPC records, planning data, flood and other peril data, and commercial sources.
- Solve entity resolution — reliably matching records across datasets to the right physical building and address.
- Analyse and model risk at building level — identifying the features that actually drive claims and turning them into signals for pricing and underwriting.
- Build, test and validate models, and clearly communicate what they do and don't tell us.
- Work with large databases and keep them accurate and performant as they grow.
- Help define what "good" looks like — data quality, coverage, model performance — and build the checks to enforce it.
Why join
- Real ownership — you'll shape a core company asset and the models built on it, not maintain someone else's legacy system.
- Equity — share in what we're building.
- Direct impact — your work shapes how we price risk and what we tell clients.
- Learn fast — work directly with the founder and a small, senior team.