A selection of the applied work behind Astravector Studio — research, internships, and independent builds across AI, geospatial analysis, and full-stack development. Each project reflects the same standard of work you'd get on a client engagement.
Sub-2-second performance for 10-node itineraries
A spatial routing engine that combines multi-source data, real constraints, and a conversational interface for itinerary planning.
Read case studySpatial risk mapping across two district datasets
A health GIS and ML pipeline for detecting patterns in dengue cases and supporting spatial public-health decisions.
Read case studyRMSE 0.07 across a cloud-cover prediction workflow
A crop spectral library and SAR-based Random Forest model that estimates NDVI when optical imagery is affected by cloud cover.
Read case studyRisk zones mapped from 7 hydrogeological parameters
A DRASTIC-based groundwater vulnerability assessment for Tirunelveli, led by a three-member team.
Read case studyIoU > 0.85 across 12 segments
A prompt-engineered Segment Anything Model pipeline that detects coastal structures from 0.5 m satellite imagery.
Read case studyRainfall-induced landslide detection in vegetated terrain
A synergistic remote-sensing framework that combines Sentinel-1 coherence with Sentinel-2 optical data.
Read case studyInteractive digitization with measurement and GeoJSON export
A browser-based tool for digitizing spatial features, measuring area and distance, and exporting GeoJSON.
Read case studyComparative crop-stage classification from InSAR descriptors
An InSAR and polarimetric-parameter study for estimating crop height and classifying paddy growth stages.
Read case studyTell us what you're building and we'll help scope the right technical path.
Start a Project