Prateek Saha

I’m an aspiring software engineer focused on backend development and data engineering, with a strong foundation in English, Python, and SQL. I enjoy turning messy data into meaningful insights through projects and hands-on experimentation.

In a recent LLM-focused intern role, I built backend features like authentication and vector search, and supported AI/ML training work with 3D LiDAR datasets. I also designed a secure multi-user architecture using JWT and Supabase Row Level Security (RLS), plus QA/data verification to maintain dataset integrity.

I’ve developed end-to-end systems that blend modern web backends with data workflows—most notably a full-stack RAG application using React, FastAPI, Supabase, FAISS, and Cohere LLM. I also worked on analytics and personalization features such as Q&A, quiz generation, and personalized learning pathways.

On the engineering side, I enjoy scalable design and continuous learning in cloud and big data practices. My object detection experience used YOLOv8, OpenCV, and Streamlit, where I improved performance through preprocessing, class imbalance fixes, and CPU-friendly inference.