AI Engineering Intern at PowersMy.biz (Live Platform)

Built a production AI assessment platform for JEE and NEET coaching institutes using FastAPI, LangGraph, React, Flutter, MongoDB, Redis, and GCP, supporting 1,000+ concurrent exam-takers. Optimized high-concurrency exam APIs using Redis caching, atomic MongoDB operations, and off-thread bcrypt processing to sustain peak load. Engineered a multimodal AI evaluation pipeline to grade handwritten subjective answers across 8 question formats and identify learning gaps. Designed an asynchronous grading queue across two services so objective scores return instantly while written-submission evaluation runs in the background. Collaborated across multiple production repositories (backend, web, mobile, and AI-agent) by raising PRs, reviewing code, resolving merge conflicts, and writing test cases. Led deployment of the FastAPI backend on Google Cloud Platform via Cloud Run with VPC connector, autoscaling, and concurrency tuning for exam-day traffic spikes.