Aditya Saini

I've built production-style AI applications spanning customer churn prediction, real-time face recognition, and Retrieval-Augmented Generation chatbots.

For a Customer Churn Intelligence Platform, I developed an end-to-end machine learning pipeline comparing Logistic Regression, Decision Tree, Random Forest, and XGBoost. The best model achieved 78.46% accuracy and a 0.57 F1 score, with FastAPI prediction endpoints and a Streamlit analytics dashboard.

I built a Smart Face Attendance System using InsightFace and OpenCV that achieved 96–98% recognition accuracy for registered users under varying lighting conditions. I also created APIs for registration, attendance tracking, and live recognition, backed by SQLite and a multi-page Streamlit dashboard.

I've also developed a PDF RAG chatbot using LangChain, Hugging Face, and ChromaDB for context-aware question answering over uploaded documents. Alongside these projects, I completed a Java development internship building four OOP-based game applications.