React + TypeScript, ASP.NET Core (.NET 8), SQL Server, EF Core, Postman ➤ Developed a multi-tenant inventory platform (ASP.NET Core, SQL Server) using SOLID principles, supporting secure tenant isolation and RBAC across inventory, HR, and accounting modules — handling 10,000+ SKUs per tenant. ➤ Engineered the data layer with EF Core (Code-First), normalized schemas, tenant-scoped query isolation, and stored procedures/views for performance-critical workflows — improving query throughput by 50%. ➤ Built a scalable React + TypeScript frontend with Context API, server-side pagination, debounced search, and secure token-based API integration — reducing UI bundle size by 35% through modular feature-driven architecture. ➤ Implemented DTO mapping (AutoMapper), structured logging, environment-based configuration, and Postman API test suites — ensuring production-grade security, maintainability, and zero-downtime deployments.

React.js, ASP.NET Core (C#), Node JS, Python, Firebase, TensorFlow/Keras, OpenCV ➤ Built a real-time indoor surveillance system using TensorFlow/Keras deep learning models, detecting violence, fire, smoke, and weapons across live webcam feeds with a classification accuracy exceeding 90%. ➤ Developed a secure ASP.NET Core (C#) RESTful API with Firebase Authentication (RBAC, admin segregation, protected routing) and Firebase Realtime Database — supporting low-latency cloud persistence at scale. ➤ Designed an interactive React.js monitoring dashboard with real-time event feeds, historical tracking, and multi-entity CRUD for users, activities, and roles.

Python, OpenAI(Ada-002/GPT-3.5), OpenCV, Tesseract, Milvus, Streamlit, Docker ➤ Engineered an AI-powered resume screening system ranking top-K candidates using cosine similarity over high-dimensional semantic embeddings — reducing manual HR screening time by 70%. ➤ Built OCR pipelines (OpenCV + Tesseract) for text extraction from image-based resumes, followed by NLP preprocessing (tokenization, lemmatization, stop-word removal), achieving 95%+ text extraction accuracy. ➤ Generated semantic embeddings (OpenAI Ada-002, GPT-3.5), stored vectors in Milvus Vector DB, and performed real-time similarity search — processing and ranking 100+ resumes per query in under 3 seconds. ➤ Delivered an interactive Streamlit HR dashboard and containerized Python microservices with Docker, enabling scalable one-command production deployment.