Projects / Overview
AI & ML Battery Health Prediction System
Engineered an award-winning (2nd place) battery health prediction platform that uses Scikit-learn regression models to forecast multi-cell battery State-of-Health (SoH) from raw telemetry. Integrated a scalable data-preprocessing pipeline with a React Native mobile interface and a Gemini-powered AI diagnostic assistant for real-time system monitoring.
PythonScikit-learnReact Native
Overview
Designed an end-to-end machine learning and mobile diagnostics platform to evaluate and predict battery degradation using multi-cell sensor data. The project placed 2nd in a university-wide competition by pairing predictive ML feature engineering with an intuitive, LLM-assisted mobile interface.
Key Features & Engineering
- Predictive ML Pipeline: Built data-cleaning and feature-engineering pipelines using Pandas, NumPy, and Scikit-learn to process complex multi-cell battery telemetry.
- State-of-Health (SoH) Modeling: Trained, evaluated, and optimized regression models for accurate battery health forecasting, maximizing performance through hyperparameter tuning.
- Real-Time Diagnostics Workflows: Developed scalable feature-aggregation pipelines to handle structured sensor streams for rapid diagnostic inference.
- Mobile App & AI Assistant: Integrated ML predictions into a React Native application featuring a Gemini-powered assistant to give users interactive, plain-language diagnostic insights.
- Competition Recognition: Awarded 2nd place in a university-wide engineering competition for system architecture, predictive accuracy, and product design.