/**
* @class ManushPatel
* @description Full-Stack Engineer bridging modern web architectures with deep learning models.
*/
class ManushPatel {
name: string = "Manush Patel";
role: string = "Full Stack Engineer & AI/ML Explorer";
education: string = "B.Tech CSE @ IIIT Vadodara (2024-2028)";
skills = {
languages: ["C/C++", "Python", "JavaScript", "TypeScript"],
frontend: ["React.js", "Next.js", "Tailwind CSS", "Socket.IO"],
backend: ["Node.js", "Express.js", "MongoDB", "MySQL", "Prisma"],
ai_ml: ["PyTorch", "TensorFlow", "OpenCV", "LangChain", "RAG"]
};
getExperience() {
return {
"Spenta Engineers": {
role: "Full Stack Developer",
period: "Dec 2025 - June 2026",
details: "Led end-to-end development of the official website from scratch. Designed UI/UX and deployed responsive architecture for a global audience."
},
"Nav Astitva Foundation": {
role: "Full Stack Developer Intern",
period: "Dec 2025 - May 2026",
details: "Engineered a scalable E-Certificate Distribution Portal using MERN stack with automated workflows."
}
};
}
getAchievements() {
return [
"2nd Runner-Up at Intellify 4.0 (TruthLens out of 200+ teams)",
"Hackatron Finalist at ABV-IIITM Gwalior (Traff-IQ out of 1000+ teams)",
"3-Star CodeChef | Pupil Codeforces | 300+ LeetCode DSA",
"98.81 percentile in JEE Main 2024 (Top 1.2% of 1.1M candidates)"
];
}
}
Dec 2025 - June 2026
- Scope: Led end-to-end development of the official corporate website for a precision drilling tools manufacturer from scratch, serving a global B2B audience.
- Architecture: Built on Next.js 16 (App Router) with Server-Side Rendering, dynamic metadata generation, and automated JSON-LD schema for SEO indexing.
- Server Actions: Implemented Nodemailer contact forms processed entirely on the server edge using Next.js Server Actions, ensuring 100% submission reliability on restricted corporate networks.
- Results: Achieved perfect 100/100 Lighthouse scores across Performance, SEO, and Accessibility. Decreased Google indexing time from 2 weeks to 48 hours.
- Stack:
Next.js 16,TypeScript,Tailwind CSS,Nodemailer,Framer Motion
Dec 2025 - May 2026
- Scope: Engineered a scalable E-Certificate Distribution Portal and digital platform for the NGO to track initiatives, publish audit reports, and streamline donations.
- Performance: Implemented React Router lazy loading and Vite code-splitting to chunk the application bundle, achieving a 65% faster initial page load (under 1.2s).
- Impact: Served 50+ multi-page audit reports with 99% reduction in client-side memory bloat. Boosted donation checkout completion rates by 25% with a streamlined 2-click payment UI.
- Stack:
React 19,Vite,Tailwind CSS,MERN Stack
TruthLens ━ Multi-Modal Disinformation & Deepfake Tracker 
- Problem: Combating the spread of AI-generated disinformation and deepfakes across news and social media, where no single detection method is sufficient.
- 4-Stage Pipeline: Engineered a multi-modal verification engine fusing Ollama LLM semantic analysis, Tavily API live web corroboration, and an Xception + Grad-CAM CNN deepfake detector for image forensics.
- Scoring Engine: Outputs a 0-100 credibility score with an explainable REAL/FAKE verdict, providing granular transparency into each verification stage's confidence.
- Multi-Platform Deployment: Shipped as a Next.js interactive dashboard, a Chrome Extension (Manifest V3) for in-browser verification, and a CLI tool for batch processing.
Traff-IQ ━ AI-Powered Adaptive Traffic Management System 
- Problem: Traditional traffic light systems are completely static, causing catastrophic delays for emergency vehicles (ambulances, fire engines) at congested intersections.
- Vision AI: Implemented YOLOv8 + OpenCV pipelines processing live video feeds for real-time vehicle detection, density estimation, and traffic violation detection (red-light jumping, overspeeding) at 30 FPS.
- Dynamic Optimization: Algorithmically optimizes traffic signal timings based on real-time vehicle density analysis, with rule-based emergency vehicle prioritization that dynamically reroutes green corridors.
- Spatial Mapping: Integrated Leaflet via react-leaflet for live geospatial tracking, rendering 10,000+ dynamic markers simultaneously with zero UI thread blocking.
- Real-Time Dashboard: Developed a React.js + Socket.IO operator dashboard with sub-200ms WebSocket latency for live event monitoring and density visualization.
AssetFlow ━ AI-Powered IT Asset Management Platform 
- Problem: IT departments lack intelligent tools to track hardware lifecycles, predict maintenance windows, and provide contextual answers about their asset inventory at scale.
- RAG Architecture: Engineered a Retrieval-Augmented Generation pipeline using Gemini Embeddings + FAISS for high-speed semantic retrieval across thousands of asset records, with LLaMA as the generative model for natural language responses.
- AI Assistant: Provides grounded, context-aware answers for asset tracking, allocation history, and predictive maintenance scheduling — eliminating manual spreadsheet lookups.
- Full-Stack Core: Built with React, Node.js, Express, Prisma, and SQL, featuring JWT authentication, Role-Based Access Control (RBAC), and comprehensive audit logging.
Acadmate ━ Unified Academic Management Platform 
- Problem: Educational institutions often struggle with fragmented systems for attendance, scheduling, and resource distribution, creating friction for both educators and students.
- Cross-Platform Architecture: Engineered a responsive, natively compiled mobile application for both Android and iOS utilizing Flutter and Dart.
- Real-Time Infrastructure: Integrated Firebase and Firestore to handle live data synchronization, robust database migration, and strict security rules across complex user roles (Students, Teachers, and Class Representatives).
- Core Modules: Developed end-to-end features including dynamic attendance tracking, PDF course material distribution, event scheduling, result publishing, and instant notifications.
SkillBuddy ━ AI-Powered Personalized Learning Platform 
- Problem: Generic online courses suffer from extremely low completion rates (~15%) because they are not tailored to individual learning paces or skill gaps.
- AI Engine: Integrated Google Gemini 1.5 Flash on the Express backend with strict JSON schema enforcement and middleware validation layers to dynamically generate personalized, 4-week learning paths in under 4.5 seconds.
- Progress Analytics: Built interactive Recharts visualizations with gamified Framer Motion transitions, increasing user session length by 40%.
- Auth & Security: Implemented stateless JWT authentication with bcrypt hashing, handling 500+ concurrent API requests with 99.9% uptime.
