For my Swiggy Delivery Time Prediction project, I developed a model using 45,000+ delivery records. The end-to-end pipeline included preprocessing, feature engineering, outlier detection, and model evaluation, and the Random Forest model achieved an R² score of 0.833.
I also implemented a Netflix movie recommendation system using collaborative filtering. It generated Top-5 personalized recommendations from user-item interactions and cosine similarity.
For my Product Management REST API project, I developed a Python and FastAPI backend for product records, with CRUD operations and MySQL storage through SQLAlchemy. I used FastAPI's Swagger UI to test and document the endpoints.
During my paid internship as a Post-LLM Training Data Annotator at Ethara AI, I contributed to data annotation, validation, and evaluation. I reviewed 1,000+ AI-generated outputs and curated datasets to support model training and continuous improvement.
