01Featured
A full-stack e-learning platform where students can browse, buy, and take courses online, and admins can manage everything from a separate dashboard. LearnEx has three parts: a student-facing site, an admin dashboard (deployed as a separate application), and a backend server.
Engineering highlights
- Course creation wizard with support for videos and images via Cloudinary
- Secure payment processing integrated through Stripe SDK
- Live notifications powered by Socket.io for real-time updates
- Analytics dashboard showing user, order, and revenue trends over 12 months using Chart.js
- CMS-like admin panel to update homepage hero, FAQs, and categories without touching code
- Multi-provider authentication with email, Google, and GitHub (NextAuth + JWT)
- Redis caching layer for API responses to improve performance
- Centralized error handling so failed requests don't crash the app
- Auth tokens stored in HTTP-only cookies with bcrypt password hashing for security
Built with
Next.jsTypeScriptTailwind CSSRedux ToolkitNode.jsExpressMongoDBMongooseRedisSocket.ioStripeCloudinaryJWTChart.js
02
CampusGuard
CompletedSep 2025 - Jun 2026
A smart campus safety and inquiry management platform engineered as my final-year capstone (awarded A+). Built around a custom 9,000+ image dataset and a 91% accurate hybrid ResNet50 + CBAM + custom head computer vision model, backed by FastAPI, AWS RDS PostgreSQL, Amazon S3, and role-based workflows for Admins, Guards, and CSA.
Engineering highlights
- Built and labeled a custom university suspicious activity dataset of 9,000+ images across 7 activity classes (2 research papers currently under review)
- Trained a hybrid ResNet50 + CBAM + custom head deep learning architecture achieving 91% accuracy for automated incident detection
- Architected a complete digital inquiry management system replacing manual paper workflows with role-based access for Admin, Security Guards, and CSA
- Engineered end-to-end incident pipeline spanning automated detection, instant real-time alerts, guard dispatch, evidence collection, and CSA departmental inquiry tracking
- Built FastAPI backend service integrated with AWS RDS PostgreSQL for relational case histories and Amazon S3 for secure snapshot storage
Built with
ReactFastAPIPythonResNet50 + CBAM + Custom HeadComputer VisionDeep LearningPostgreSQLAWS RDSAmazon S3Tailwind CSSREST APIs
03
Vendora
CompletedFeb 2026 - Apr 2026
A production-grade multi-vendor e-commerce marketplace connecting customers, independent sellers, and platform administrators across four distinct deployment units with real-time buyer-seller communication and multi-gateway payment processing.
Engineering highlights
- Architected a decentralized multi-vendor marketplace across 4 deployment units (Vercel frontend, Render Express API, Render Socket server, MongoDB Atlas)
- Built real-time buyer-seller chat system with message delivery and read indicators using a dedicated Socket.io server
- Engineered dual payment workflows integrating Stripe and PayPal SDKs with automated seller commission and withdrawal calculations
- Designed three specialized role-based interfaces for buyers, autonomous vendor storefronts, and administrative governance
- Implemented promotional mechanisms including timed flash sale events, coupon codes, and Cloudinary-optimized media delivery
Built with
React 19ViteTailwind CSSMaterial-UIRedux ToolkitNode.jsExpress.jsMongoDB AtlasMongooseSocket.ioStripePayPalCloudinaryJWT
04
Mockly
CompletedOct 2025 - Present
A full-stack AI-powered technical interview platform that generates contextual interview questions, enables voice-based answers via speech-to-text, and delivers automated AI evaluation with detailed analytics.
Engineering highlights
- Integrated Google Gemini AI (gemini-2.5-flash) for contextual technical question generation and automated performance evaluation
- Integrated native Web Speech API for real-time speech-to-text voice answer input
- Implemented OTP email verification, JWT authentication with HTTP-only cookies, and bcrypt password hashing
- Built comprehensive performance analytics dashboards with score tracking and interview history using Recharts
- Applied rate-limiting caps (3 interviews/day) on RESTful endpoints to optimize AI token consumption and prevent abuse
Built with
React 19ViteTailwind CSSRedux ToolkitNode.jsExpress.jsMongoDBMongooseGoogle Gemini APIWeb Speech APIJWTNodemailerRecharts