Data Scientist / Data Analyst
hace 1 día
Birmingham
Role: Data Scientist / Data Analyst — Property Use-Case Modelling Company: adema.ai (UK PropTech) Location: Remote-friendly (UK/EU time zones) • Full-time Mission Help us add new property use-case analyses (e.g., Residential, Social Housing, Commercial, Care, STR, EV Charging, Data Centres). You’ll research and source datasets, build models that infer demand/supply and revenue potential at the most local level possible, and ship them into our product. What you’ll do • Map each “use case” data landscape: Identify, evaluate and acquire structured sources (e.g., prices, rents, demographics, planning, POIs, transport, connectivity) and useful unstructured sources (local plans, market reports, PDFs). Track licence terms and provenance., • Engineer geospatial & temporal features: Join/clean data, spatially downscale/coalesce (e.g., LA → LSOA/sector/property) using proxies (prices, comps, time-series trends, neighbourhood features, travel times)., • Build predictive/forecast models: Estimate demand, supply, pricing/rent & revenue; quantify uncertainty; design robust validation and back-testing., • Productionise your work: Persist outputs in Postgres/PostGIS, expose via GraphQL; implement services in Go or Python; write clear SQL views, tests and docs; monitor data quality and model drift., • Extract signal from unstructured data: Scrape/download reports, parse tables/figures, apply LLM-assisted extraction where useful; convert to structured features., • Collaborate across the stack:, • With Product to define success metrics and MVP scope per genre., • With Backend to integrate pipelines/APIs., • With Frontend/AI teams to shape GraphQL queries and agent/tool schemas., • Ship iteratively: Prioritise “easier” genres first (Residential, Commercial), then expand to specialised sectors. Document assumptions and limitations. What you’ve done • 3+ years in Data Science / Analytics (or 2+ with a strong portfolio) delivering models into production., • Strong Python (pandas/numpy/scikit-learn; XGBoost/LightGBM; basic PyTorch a plus) and SQL., • Solid geospatial skills (PostGIS/GeoPandas/QGIS) and time-series/forecasting know-how., • ETL/ELT and data wrangling at scale; comfort with scraping and PDF/table extraction., • Good software practice: Git, containers, CI/CD, testing, clear documentation., • Product mindset: bias to ship, explain results simply, track impact. Nice to have • Go, GraphQL, dbt, Airflow/Dagster, FastAPI; Azure., • UK property/economics exposure (Land Registry, EPC, census/ONS, planning, VOA etc.)., • LLM/AI experience for information extraction or analyst co-pilots. Success looks like 30 days: Different types of Residential models live in app (Postgres/PostGIS + GraphQL), with documented features, validation and property-level scoring. Why adema.ai We’re building the decision layer for UK property — rigorous data, clear modelling, and real-world utility. If you love turning messy datasets into decisive answers, you’ll fit right in. HOW TO APPLY Please send your CV to and we will come back to you quickly.