Mantavya Patel
I am an engineering student currently pursuing a Bachelor of Technology (B.Tech.) in Computer Science and Business Systems. I am passionate about AI / ML, Software and Full-Stack Development—and yes, I spend an unhealthy amount of time wondering why the code worked five minutes ago.
[2024 – Present] I am pursuing my B.Tech at VIT-AP University, Andhra Pradesh.
[2023 – 2024] I completed my higher secondary education (12th cbse boards) at Shree Swaminarayan International Gunatit Vidhyadham, Gujarat.
[2009 – 2023] I completed my school education at Shree Shreeji English Medium School, Gujarat.
Top 106 Globally in Google Solution Challenge 2026
Github Repository Video PPT
Problem statement : Rapid Crisis - Smart Fire Evacuation
Rapid Crisis is an end-to-end smart fire evacuation system that combines a React dashboard, cloud backend, and Raspberry Pi edge controller. It receives fire alerts and sensor data from ESP32 devices, detects hazards, calculates evacuation routes, and guides people using LED indicators. The system stores building layouts and telemetry in Firestore and sends real-time updates to the dashboard using Server-Sent Events (SSE). It also includes a simulation mode for testing fire spread and evacuation without physical hardware.
SIH 2026 - Shortlisted in university internal rounds among 564 team for finals
Github Repository Youtube_Link PPT
Problem statement : BioTech - AI-Powered Dynamic Mental Health Monitoring and Distress Prediction System for Victims of Atrocities
Mann Shathi is a multimodal mental health platform that uses Text, Voice, and Biosignal analysis to detect distress, emotional dissonance, and changes in a victim’s baseline. It generates an empathetic AI response, calculates a unified distress score, securely stores trends in Supabase, and alerts counsellors with prioritized cases and relevant legal/relief recommendations.
Ai Ad Generator
Github Repository Live Deployed Report
AiAdGenerator is a web application that transforms product and model images into AI-generated advertisements, including images and videos. Users provide project details, upload assets, and generate ads through Google GenAI.
The platform uses Clerk for authentication, Cloudinary for media storage, and PostgreSQL with Prisma for database management, with generated projects stored for later viewing and management.
VIT - AP Bookify
Github Repository Live Deployed Report
Bookify is a full-stack web application where users can sign in, create and browse book listings, and comment on listings. Authentication is handled through Clerk, while book data and comments are stored in PostgreSQL using Drizzle ORM.
User Authentication — Secure sign-in using Clerk.
Book Listings/Comments - Users can create and browse book listings + comment.
Backend API — Express + TypeScript handles application requests.
Database — PostgreSQL stores books, comments, and user-related data.
ORM — Drizzle ORM manages database operations.
Saaransh AI
Github Repository Live Deployed Report
Saaransh-AI (Text Summarizer) is a lightweight Generative AI application that converts dialogues and text into concise abstractive summaries using a fine-tuned T5 Transformer trained on the SAMSum dataset.
T5 Transformer — Fine-tuned sequence-to-sequence model for abstractive summarization.
SAMSum Dataset — Trained on real-world dialogue and human-written summaries.
FastAPI Backend — Handles text processing and model inference.
Hugging Face Hub — Hosts the fine-tuned model for inference.
Languages: Python, C++, JavaScript, TypeScript, SQL
AI/ML: Scikit-learn, PyTorch, Hugging Face, Reinforcement Learning, LLM, RAG, GAN, Agentic AI, End-to-End ML/DL Pipeline
Frontend: React, Tailwind CSS, JavaScript, PWA, DaisyUI, TanStack Query
Backend & DevOps: Flask, FastAPI, Node.js, Express.js, REST APIs, Docker, Git
Databases & Storage: PostgreSQL, MongoDB, Firebase, Prisma, Drizzle
Core Fundamentals: Data Structures & Algorithms, OOP, DBMS, Operating Systems, System Design
