Reading the planet in patterns.

Presents an end-to-end spatio-temporal framework coupling Sentinel-5P TROPOMI satellite imagery with sparse ground IoT air sensors across central Indian urban corridors. Formulates dynamic adjacency matrices driven by boundary-layer wind advection, demonstrating a 23.4% reduction in root mean square error during extreme winter thermal inversion events.

Investigates multi-sensor fusion combining high-revisit Moderate Resolution Imaging Spectroradiometer (MODIS) Aerosol Optical Depth (AOD) with micro-scale meteorological variables to resolve continuous ground-level PM2.5 concentrations over complex topographical terrain.

Benchmarking classic ensemble techniques against deep recurrent and attention models for multi-horizon urban air quality prediction. Identified seasonal lag dependencies and validated feature attribution models to enhance public environmental transparency.

Synthesizes two years of academic classroom leadership mentoring university undergraduates across Database Management Systems, Data Mining, and Operating Systems. Evaluates interactive software scaffolds to accelerate student mastery of abstract algorithmic structures.