GIS analyst · Geospatial engineer · Turin
I build WebGIS tools, spatial data pipelines and GeoAI models for infrastructure and natural-hazard work, from the raw data to the map someone has to make a decision with.
01 Selected work
Input ortho
Under the lens: height (DSM)
GeoAI · Scientific Reports (Nature) · 2024
DeepLabv3 and U-Net segmentation with RGB and height (DSM) fusion, benchmarked across five architectures on very-high-resolution imagery. Move over the image: the lens shows the height layer the fused models also read.
Natural hazards · Sustainability (MDPI) · 2023
A multi-hazard road network analysis for the Orco Valley: flood and landslide scenarios, service areas for hospitals and public facilities, and origin-destination travel times, on a PostGIS database built from open 2D and 3D data. The terrain at the top of this page is the same valley.
WebGIS · Mapbox GL JS · Client work
A browser map for reviewing parcels without leaving the map: custom basemap, polygon-draw selection, attribute filters, notes and CSV/KML export, built directly on Mapbox GL JS.
Rail infrastructure · Flagship prototype
Rail monitoring that puts GIS layers, InSAR deformation, UAV inspection and point-cloud review in one decision surface, with audit logging and report export.
02 Research
Citations from ResearchGate, September 2026. h-index from Google Scholar.
Describing what spatial datasets mean, encoding it as metadata, and turning plain-language questions into spatial operations a system can run and defend.
03 Open-source tools
Simplifies infrastructure line networks while pinning shared vertices, so junctions never disconnect.
Finds every line-line intersection and builds a planar edge-node graph. STRtree backend, 2.5× faster on large networks.
Simplifies polygon coverages arc by arc so neighbours stay identical. No gaps, no slivers.
Tile-streamed slope, aspect, hillshade and focal stats on Cloud-Optimized GeoTIFFs, with no seams at tile edges.
Monte Carlo zonal difference of ecosystem-service value between land-cover scenarios, with credible intervals.
Portable JSON sidecar indexes for LAS/LAZ point clouds, with a browser preview of bbox queries.
04 About
I studied civil engineering at NUST in Pakistan and did my MSc in infrastructure engineering at Politecnico di Torino, where my thesis used machine learning for corridor mapping.
Since then I have worked on hazard and resilience research with Politecnico di Torino and DigiSky, alongside freelance GIS work. On Upwork I was Top Rated for three years, with more than 550 projects for clients that include the WHO and oil and gas operators in the US and Canada.
I'm most useful where analysis meets software: a PostGIS schema, a QGIS plugin, a segmentation model or a WebGIS front end, built so someone else can check and re-run it.
World Health Organization · Politecnico di Torino (DIATI) · DigiSky · ESA InCubed · NODES · MANAGE 5.0
05 Contact
Based in Turin (CET). Open to GIS roles, research positions and freelance projects, remote or on-site in the EU.
daudtasleem215@gmail.com