Geospatial Data Science Engineer

About the role

Unikie is looking for a Geospatial Data Science Engineer to build the data layer behind its next-generation software-defined electric vehicles. You will turn massive, high-frequency streams of vehicle sensor observations into accurate, dynamic map features used for localization, routing and ADAS. The role calls for advanced computational statistics, machine learning and spatial analysis, applied to large datasets drawn from fleet telemetry.

Key responsibilities

• Design, build and optimise scalable pipelines that ingest, clean and segment billions of daily sensor observations into map features and geometries.
• Apply data mining and analytics to detect real-world changes such as road closures, new lanes and construction, and design statistical tests to validate map accuracy.
• Develop algorithms for map matching, trajectory smoothing and sensor fusion to turn noisy probe data into high-fidelity road geometry and attributes.
• Lead the technical design of spatial data systems, with low-latency queries, efficient spatial indexing (e.g. H3, S2) and scalable storage of vector and raster map data.
• Work with embedded software engineers, perception teams and cloud architects to align onboard vehicle capabilities with cloud-based mapping platforms.
• Stay up to date with big data, computational statistics, machine learning and automated mapping advances.

Required profile

• Bachelor's degree in Statistics, Applied Mathematics, Computer Science, Geoinformatics, Robotics or a related quantitative field, with work on computational statistics, data mining, machine learning or spatial optimisation.
• 8+ years of experience building and maintaining large-scale data processing and predictive systems.
• Expert knowledge of data mining methods such as regression, classifiers, clustering, association rules, decision trees and Bayesian networks.
• Proven experience with geospatial data, noisy time-series datasets, PostGIS queries and spatial libraries/indexing systems (H3, S2, GeoPandas).
• Strong statistical foundation in designing, running and analysing complex data tests and validation.

Preferred qualifications

• Crowdsourced map generation experience (probe-data aggregation, change detection, incremental updates).
• Cloud deployment of geospatial pipelines and ML workloads (AWS, GCP, Databricks).
• Open-source tools such as Martin, QGIS, GDAL/OGR, GEOS; automotive map standards such as NDS, OSM, Lanelet2/HD Maps.
• Familiarity with robotics concepts such as SLAM, sensor fusion and point cloud processing.