Mohamed Amine Hajji / Experience

Full-Stack Software Engineer — Part-Time — DeepVolt

Owned the geospatial data-to-pixels path of a location-intelligence product: from raw GeoJSON to vector tiles to interactive Mapbox UIs, plus the filters and datasets that feed the site-scoring algorithm.

• Designed backend API endpoints (Node.js/Express) for processing and serving large-scale spatial datasets.
• Supported the data science team with data cleaning and web scraping to build training datasets, using QGIS to visualize and validate geospatial data.
• Deployed and maintained services on AWS (EC2, RDS, S3).

Geospatial vector tiling pipeline

From multi-GB GeoJSON to sub-second interactive maps: a full tiling pipeline with dynamic, filter-driven clustering.

• Tippecanoe tiling strategy tuned per zoom level (simplification, feature dropping, cluster radii), hosted on S3.
• Led adoption of Martin (Rust/PostGIS tile server): wrote the PostGIS queries and deployed it for real-time, filter-driven clustering.
• Interactive Mapbox GL visualizations in Angular; PostgreSQL storage; deployed on AWS (EC2, RDS).

Site-scoring filters & market expansion

Configurable competitor/POI/traffic filters for the site-scoring algorithm, a 100k+ EV-station dataset, and the front-end for DeepVolt’s Istanbul expansion.

• Configurable competitor, POI, and traffic filters driving the site-scoring algorithm.
• Optimized TomTom API data fetching; integrated a 100,000+ station EV-charging dataset (France).
• Front-end for the Istanbul expansion: AI-recommended charging-station locations wired to new backend endpoints.

Stack: Angular, TypeScript, Mapbox GL, Tippecanoe, Martin, TomTom API, Node.js, Express, PostgreSQL, QGIS, AWS