Developed a drug-drug and drug-food interaction engine using RxNorm data and custom CSVs, complete with severity scoring and alternative medication suggestions.
Designed the medication alternative recommendation module, mapping brand names to generics for consistent standardization.
Built a Flask web application for uploading medical documents, reviewing extracted medications, and running interaction checks via an intuitive UI.
Used Azure Text Analytics service to extract and normalize medication entities from unstructured clinical text for downstream processing.
A Survey of Adversarial Attack and Defence Methods
Conducted a survey of adversarial attack and defense techniques for deep learning.
Categorized methods across multimedia, autonomous systems, and finance.
Predictive Maintenance in Electrical Systems
Developed a hybrid Bi-LSTM + Random Forest framework for predictive maintenance.
GreenGrids – Bangalore LST Analysis Tool
Wrote Python scripts leveraging the Google Earth Engine API to collect and preprocess Landsat 8 LST data for Bengaluru’s neighborhoods.
Processed and converted OpenStreetMap boundary data and auxiliary environmental layers into web-ready GeoJSON format.
Overlaid LST heatmap and mean LST choropleth layers onto the interactive website map with toggle controls.
Smart Optimized EV Public Charging System
Developed a smart EV charging system with queue management.
Engineered an anti-fraud mechanism to secure and optimize public EV charging infrastructure.
Architected a centralized multi-factor prioritization solution for managing EV charging stations.
