AI/ML Geospatial Consultant – Part Time engagement 12‑month partnership, with potential ongoing work£400 to £500 a day (Outside IR35)Remote Working Technical Focus:
Advanced AI/ML model development, geospatial analytics, ArcGIS, Python, and GIS‑enabled modelling.My client is developing a new suite of geospatial digital products powered by AI and machine learning.We’re looking for an experienced AI/ML Consultant
to partner with us to design, build, validate, and operationalise these frameworks. You’ll provide strategic and technical guidance across the full project lifecycle, supporting the Project Director, Research and Intervention Lead, and our in‑house Data Analyst. While this project is scoped for
roughly 1 day a week for 12 months, we’re seeking a consultant who can also offer ad‑hoc support as the work evolves.Experience required: -
1. Advanced expertise and proficiency in AI, predictive analytics and machine‑learning model development (minimum 5 years professional experience) and its practical application in a charity or business setting.
2. Proficiency in spatial data science and GIS‑enabled modelling (QGIS, ArcGIS Pro, or Python GIS stack).
3. Proficient with PyTorch, Ultralytics YOLO and data management on AWS or Azure.
4. Demonstrable ability to build and validate models using UK‑based socio‑environmental datasets (, IMD, ONS Open Data, land use, accessibility).
5. Experience integrating models into offline or on‑prem environments.
6. Strongmunicator able to bridge the gap between data science and decision‑making; skilled at explaining technical detail clearly and tailoring language for different audiences.
7. Capable of identifying potential risks or barriers early,municating them clearly, and proposing practical mitigation strategies.
8. Demonstrated skill in maintaining stakeholder alignment around project milestones and deliverables.
9. Prior experience in environmental risk modelling, urban analytics, or behavioural data analysis.
10. Understanding of geostatistics, spatial interpolation, and postcode‑level disaggregation.
#4758779 - Simon Iatrou