A large-scale, cloud-native Picture Archiving and Communication System (PACS) designed to manage medical imaging workflows.

• Architecture & Scale: Engineered the system to handle 150x peak traffic loads and reduced database costs by 30% via advanced query optimizations and caching strategies.
• Performance: Implemented a Backend-for-Frontend (BFF) pattern, reducing application load times by 25%.
• Security: Designed a granular Role-Based Access Control (RBAC) system for strict, secure data isolation in a multi-tenant environment.
• AI Integration: Integrated generative AI models into core healthcare workflows to boost practitioner efficiency.

A robust C# desktop application for X-ray acquisition, serving as the bridge between hardware devices and the VPAX cloud ecosystem.

• Hardware Integration: Enhanced compatibility across diverse client setups by integrating new X-ray devices and sensors.
• Seamless Sync: Developed secure data pipelines to ensure real-time synchronization of heavy imaging data with the VPAX cloud.
• Core Development: Maintained the full lifecycle of the application, ensuring stability and performance in clinical settings.

A next-generation email client built to streamline inbox management using Generative AI.

• RAG Implementation: Built a "Chat with your Inbox" feature using Retrieval-Augmented Generation (RAG) and LangChain.
• AI Features: Implemented automatic thread summarization, daily digests, and context-aware reply generation.
• Architecture: Designed a provider-agnostic backend with NestJS and Postgres, allowing seamless integration with Gmail, Outlook, and other IMAP services.

A scalable, event-driven e-commerce exchange that allows users to browse, buy, and sell event tickets.

• Microservices Architecture: orchestrated using Kubernetes and Docker for high availability.
• Event-Driven Design: Utilized NATS Streaming Server for asynchronous service communication and data consistency.
• Tech Stack: Built with NodeJS/Express (Backend), NextJS (Frontend), Redis (Caching), and Stripe (Payments).
• Reliability: Secured with JWT authentication and fully tested with Jest and CI/CD pipelines.

A research-grade 3DSlicer plugin for cardiac CT assessment and health risk calculation.

• Deep Learning: Uses PyTorch models for accurate segmentation of coronary artery calcification.
• Hybrid Processing: Offloads heavy computation to an AWS EC2 Flask server while maintaining a responsive local PyQt5 interface.
• Impact: Results and methodology were published in the 2024 41st National Radio Science Conference (NRSC).

A comprehensive management system designed for student organizations, facilitating recruitment and operations.

• Impact: Deployed and actively used by 2,000+ students.
• Key Features: Manages recruitment pipelines, interview scheduling, member databases, and KPI monitoring.
• Integration: Seamlessly syncs data with Google Sheets via Google APIs for non-technical admins.

A Computerized Maintenance Management System (CMMS) tailored for medical facilities.

• Asset Management: Tracks inventory and manages lifecycles for medical equipment.
• Workflow: Handles preventive and corrective maintenance work orders.
• Real-Time: Features live notifications for urgent maintenance requests using Socket.IO.

A secure, production-ready REST API designed to automate retail vending operations.

• Security: Implements JWT-based Role-Based Access Control (RBAC) for admin and user security.
• Logic: Handles complex transaction logic including deposits, purchases, and exact change/withdrawal calculations.
• Documentation: Fully documented using Swagger/OpenAPI specifications.

A full-stack application enabling seamless food ordering for restaurants.

• State Management: leveraged Redux for complex cart management and user state.
• Type Safety: Built entirely with TypeScript to ensure code reliability and maintainability.
• Stack: React frontend connected to an Express/MongoDB backend.

A specialized Hospital Information System (HIS) for dental clinics.

• Full-Stack Development: Designed the database schema and built both the backend (NodeJS) and frontend (Bootstrap/jQuery).
• Data Handling: Used Sequelize ORM for structured interaction with the MySQL database.

A versatile desktop tool capable of ingesting and analyzing tabular, image, and text data.

• Tabular: Calculates descriptive statistics and covariance using Pandas.
• Image: Generates color histograms and segmentation masks using OpenCV.
• NLP: Utilizes Word2Vec and T-SNE to visualize text vector embeddings.

An audio fingerprinting application similar to Shazam.

• Algorithm: Extracts audio features from source files and stores them in a fingerprint database.
• Matching: Identifies recorded songs by comparing features and calculating similarity percentages.
• UI: Built with PyQt5 for a responsive desktop experience.

A real-time audio player and signal processing tool.

• Signal Processing: Uses NumPy and SciPy for real-time EQ and analysis.
• Visuals: Animated frequency visualizations synced to playback.
• UI: PyQt5 desktop interface with responsive controls.

Dual-mode robotic vehicle: CV-based autonomous lane detection and manual MeteorJS control with live streaming.

ESP8266 + Firebase + MeteorJS WiFi triangulation to map device location indoors in real-time.

Industrial heating simulation via SPI; supports multiple logic states (Standby, Operational, Normal, Error) with ATmega32.