Full-stack digital permit platform with a 5-role RBAC approval workflow, cutting permit turnaround from 3 days to under 4 hours. Built an OCR pipeline (EasyOCR + OpenCV) for automated Aadhaar/DL data extraction at 94%+ accuracy, plus QR-based mobile exam access.

Fine-tuned YOLOv8 on 5,000+ custom-annotated baggage images, achieving 92%+ detection accuracy. Integrated perspective geometry for real-time dimensional classification and a Flask API triggering staff alerts within 1.5 seconds of detection.

Independently built 4 AI modules — PPE Detection (90%+ precision), Smoke & Fire Detection, People & Vehicle Counting (DeepSORT), and Intrusion Detection — with Flask REST APIs feeding a real-time Angular dashboard over WebSockets across 90+ live camera feeds.

FaceNet-based biometric identification pipeline (96%+ accuracy) with liveness detection; reduced per-face inference from 800ms to 120ms, supporting 500+ concurrent passenger flows with sub-2-second gate auto-identification via a FastAPI + Angular stack.

Real-time flight information portal with RxJS-driven auto-refresh every 30 seconds; Spring Boot APIs aggregate live AODB and airline feeds, serving 500+ passenger queries/hour with query performance optimised by 40% via indexing and server-side pagination.

Designed 80+ parameterised operational report views with multi-field filtering and CSV/PDF export; virtual scrolling for 10,000+ record datasets cut average report generation time from 15 minutes to under 2 minutes, reducing manual reporting effort by ~60%.

Kidney-Disease-Classification-MLflow-DVC

The project integrates MLflow, DVC, and AWS for classification and deployment. It includes tracking experiments, building pipelines, Docker-based EC2 deployment, and GitHub Actions for CI/CD automation.

Sales Performance Dashboard

The project showcases Power BI's capabilities in analyzing and visualizing sales data. It features interactive dashboards, key metrics like sales growth and regional performance, and actionable insights for data-driven decisions.

Netflix and TV Shows Clustering

This project aims to cluster Netflix content based on attributes such as Director, Cast, Country, Genre, and Description. By leveraging ML techniques.

Brain Tumor Detection

This project aims to detect brain tumors from MRI images using deep learning techniques. The goal is to provide a reliable, automated tool for early diagnosis and treatment planning.

EDA on Insurance Dataset

The EDA_on_Insurance_Dataset project involves analyzing an insurance dataset to uncover key trends and factors affecting insurance costs using data visualization and statistical summaries.

Full Stack Development

End-to-end web application development with Angular, Spring Boot, Flask and FastAPI — from responsive, reactive frontends to production-ready REST APIs and microservices.

Computer Vision & AI Solutions

Custom computer vision pipelines using YOLOv8, OpenCV, FaceNet and DeepSORT for real-time detection, tracking, and biometric identification systems.

Backend & API Development

Scalable RESTful microservices and AI inference APIs with Spring Boot, Flask and FastAPI, featuring JWT authentication, RBAC and Swagger/OpenAPI documentation.

Real-Time Dashboards

Angular dashboards with RxJS and WebSocket-driven live data binding for real-time alerts, analytics, and operational reporting at scale.