Piyush Pareek

I'm a Computer Science student building full-stack applications and AI/ML solutions that turn complex data into practical outcomes.

At MUFG (MPMS), I'm leading a POC for AI-powered document optimization to extract actionable insights from unstructured data. I've also engineered automated test-case-generation concepts and supported PEGA development and testing for distributed financial systems.

I built a DeepFake Detection System that trained on more than 200,000 images, achieved 89% prediction accuracy, reduced overfitting by 40%, and delivered real-time Streamlit predictions averaging 100ms. For Grammie, I built an AI healthcare platform with an 87% accurate predictive matching system, a drug recommender, and a JWT-secured Node.js backend.

My projects also include an e-learning platform with CRUD APIs and MySQL integration, plus Foresight, an AI-powered auto-scaling platform using Prophet, Isolation Forest, Docker, Kubernetes, React, and FastAPI. I enjoy applying data science and full-stack engineering to automation, scalable systems, and real-world challenges.