I've started working on projects this summer as I progressed my way through different tech stacks. You may check my entire work on Github.
Some of the notable work of mine:
Video Summarization using Deep LearningDesigned MambaVSum, a novel architecture replacing O(N2) self-attention with Bidirectional Selective State-Space Models
(BiMamba), achieving O(N) linear-time complexity for long-form video sequences. Achieved 56.27 F-Score on TVSum with 756K parameters (3× fewer than VASNet ) and lower cross-split variance (±2.20 vs ±3.09) Engineered a vectorized chunked cumulative product scan in pure PyTorch, delivering 20× training speedup full 5-fold training in
10 min on an RTX 3050 (4GB VRAM)
E-Commerce Web AppFull-stack MERN e-commerce platform with JWT authentication, admin roles, product CRUD, cart and order management, Stripe payments, Cloudinary uploads, and Redux-based global state management.
Phishing Detection SystemAI-powered phishing detection system using PyTorch LSTM with 94.1 F1-score, scalable Node.js REST API, Chrome extension for Gmail integration, and Google Safe Browsing API for URL validation.
Krishi Kiran – AI-Powered AgriTech PlatformFull-stack agri-commerce platform integrating crop disease detection, multilingual AI chatbot, real-time weather data, and crop price comparison using TensorFlow and modern web technologies.
Chemical Department NIT Patna – FINS WebsiteDeveloped a responsive departmental website using React and Google Apps Script with dynamic content management for notices, events, and conference updates.