Case Study / Full-Stack AI Platform
PREPMEUP AI
An autonomous, university-aligned exam preparation platform engineered for B.Tech students. Combines multimodal vision OCR, RAG grounded to exact syllabus modules, browser-native BEE circuit engines, and instantaneous WhatsApp OTP authentication.
System Architecture
Engineering Constraints & Solutions
Multimodal Vision & Curriculum RAG
Engineering students frequently ask doubts containing complex circuit diagrams or handwritten equations. PrepMeUp ingests base64 visual payloads directly into Groq Llama-3.2-11b-vision-preview for zero-latency OCR. Extracted queries are automatically cross-referenced against syllabus modules and published notes before synthesis. If out-of-syllabus, the model explicitly flags its source as general knowledge.
Zero-Dependency Browser Visualizers
Rather than relying on resource-intensive backend WebGL runtimes or remote sandboxes, 10 interactive visualizers for Basic Electrical Engineering (KVL/KCL, AC phasors, RLC resonance, Thevenin equivalents) and Python execution flow (recursion call stack, memory pointers, loop iterations) run entirely client-side using pure SVG, HTML Canvas, and React state machines.
WhatsApp OTP & Razorpay Reconciliation
Engineered frictionless passwordless onboarding tailored to Indian university students via 6-digit WhatsApp OTPs delivered through the Meta Cloud API. Monetization is secured via Razorpay Orders and HMAC-SHA256 signature verification with an atomic server counter granting the first 10 students 100% free lifetime access before locking.
Live Production Deployment
Experience PrepMeUp AI in Action
Test the multimodal doubt solver, interact with circuit simulators, and explore the complete semester syllabus.