Khan Rahat

I’m a Machine Learning Engineer and recent B.Sc. Computer Science graduate based in India, focused on turning real-world data into reliable, deployable ML systems. Through my ML Engineer OJT, I worked on an end-to-end RAG system pipeline—handling data ingestion, knowledge chunking, and content structuring for LLM-based retrieval, including preparing transcripts and organizing knowledge chunks for RAG ingestion.

I bring hands-on experience with NLP and applied ML: I built an AI-powered crime case analysis system using spaCy NER and a TF-IDF cosine similarity pipeline for related-case detection, and I developed a Flask app with a JWT-secured REST API plus MySQL storage for case management and PDF report export. I’ve also deployed multiple interactive models—an IPL match winner predictor (51.9% CV accuracy) via Gradio on Hugging Face Spaces, a crime type prediction app with real-time Streamlit outputs and Plotly confidence charts, and a house price prediction project that compared Linear Regression vs. Random Forest after EDA-driven insights on multicollinearity.