Neural Credit
NeuralCredit_ is an advanced, full-stack Fintech web application that bridges the gap between interactive React frontends and predictive Machine Learning backends. By dynamically spanning Python child processes from Node.js, it evaluates real-time personal loan applications against a pre-trained Random Forest Classifier.
Beyond simple predictions, NeuralCredit_ implements Explainable AI (XAI), enterprise-grade UX (such as Open Banking simulators and PDF generation), and automated SQL analytics dashboards.
This project demonstrates strong Architectural Engineering, bridging Data Science and Web Development into a cohesive Software-as-a-Service (SaaS) product.
Portfolio
Engineered a high-performance, cinematic developer portfolio featuring advanced browser compositor optimizations, physics-based cursor interactions, and a custom animation system built entirely with Framer Motion.
🔹 Custom Neural Cursor System: Built a multi-mode custom cursor using useMotionValue and useSpring, bypassing React's render cycle entirely to achieve lag-free, 60fps+ tracking with a decoupled ghost trail ring and dynamic morphing states (dot → beam → ring) based on hovered element type.
🔹 Magnetic Skill Grid (Arsenal): Engineered isolated physics-based "magnetic pull" animations on individual skill components using per-component useMotionValue instances, ensuring zero cross-component re-renders during rapid mouse movement.
🔹 Cinematic Intro Sequence: Designed a multi-stage boot sequence featuring terminal-style status logs, 3D grid expansion animations, and choreographed reveal transitions using Framer Motion's AnimatePresence and staggerChildren orchestration.
🔹 Ambient Project Gallery: Implemented a fixed-position, low-opacity background image preview system that responds to project hover states with smooth cross-fade transitions, maintaining a clean, editorial aesthetic without cluttering the viewport.
Tech Stack: React.js · Vite · Framer Motion (useMotionValue, useSpring, AnimatePresence) · Vanilla CSS · Responsive Design
Travelling Tent
Developed and deployed a production-ready, full-stack expedition booking platform featuring real-time geospatial mapping, secure authentication, and a complete e-commerce checkout pipeline with integrated payment gateway support.
🔹 Geospatial Integration: Implemented interactive map components using React-Leaflet, enabling users to visually explore expedition destinations, view geographic coordinates, and interact with location-based data—all rendered seamlessly within the React SPA without page reloads.
🔹 E-Commerce Architecture: Architected a robust shopping cart system with dynamic pricing calculations, quantity management, and a streamlined checkout funnel. Integrated the Razorpay Payment Gateway API for secure, production-grade transaction processing.
🔹 Authentication & Route Protection: Implemented Firebase Authentication for secure user registration and login. Engineered protected route guards using React-Router-DOM to restrict unauthorized access to cart, checkout, and user profile dashboards.
🔹 Performance & Build Optimization: Leveraged Vite as the build tool for aggressive Hot Module Replacement (HMR) and production bundle optimization. Structured dynamic URL routing (e.g., /trip/:tripId) for scalable content delivery as the expedition catalog grows.
🔹 Cinematic UI/UX: Designed a visually immersive interface featuring Framer Motion-powered page transitions, parallax scrolling effects, and responsive layouts optimized across all device breakpoints.
Tech Stack: React.js · Vite · Node.js · Firebase Authentication · React-Router-DOM · Razorpay API · React-Leaflet · Framer Motion · CSS3
Ledger AI
Architected a MERN-stack B2B automation SaaS for micro-SMEs in India. Engineered an AI-powered Snap & Reconcile feature using OpenAI Vision API to parse receipt images into structured JSON financial ledgers, and designed a real-time WebSocket-driven user experience.
NeuralCredit_
Developed an explainable fintech engine using a Python Random Forest classification model to predict creditworthiness. Built a React dashboard with dynamic radar charts, joint-applicant state-merging logic, a scalable MERN backend, and Framer Motion state transitions.
