Rhea MILEPILLI

• Measured search ranking performance with ndcg10 and mean reciprocal rank (mrr), increasing baseline score from 0.542 to 0.724.
• Engineered a search ranking engine using Python and LightGBM LambdaMART to increase search relevance.
• Built an AI research assistant by integrating Claude Projects, NotebookLM, and MCP tools.
• Configured and deployed a responsive portfolio website using HTML, CSS, GitHub Pages, and Formspree.
• Implemented prompt engineering with Git and GitHub to streamline development documentation and version control.

React, FastAPI, Python, Whisper, Ollama, Qwen 2.5

• Developed an AI-powered platform that converts uploaded videos into transcripts using OpenAI's Whisper.
• Implemented video upload, transcription, and automated content generation workflows.
• Integrated Qwen 2.5 via Ollama for generating platform-specific captions, posts, and descriptions, streamlining content creation processes.
• Developed REST API endpoints using FastAPI, and integrated frontend-backend communication using Axios.

Python, Pandas, NumPy, Scikit-learn

• Created forecasting machine learning model driven by historical sales data for sales demand.
• Executed data preprocessing, exploratory data analysis, and feature analysis to uncover sales trends.
• Measured model performance using mean absolute error (MAE) and R² score metrics.