Designed and implemented a full ETL pipeline to clean and transform health and nutrition data from multiple sources into a unified MySQL/Snowflake-style schema. The pipeline handles data ingestion, validation, transformation, and loading with emphasis on data quality and consistency.

Built a full NBA season simulation using parallel programming principles (multithreading and multiprocessing) in Python. Achieved significant reductions in execution time through efficient resource utilization and concurrent game processing. Implemented SQLite for persistent storage of simulation results and statistics.

Built a deep learning-based waste classification system using a custom CNN and MobileNetV2. Combined binary and multiclass image classification models into an integrated pipeline to improve accuracy and support automated waste sorting for sustainability initiatives. Achieved high accuracy in distinguishing between recyclable and non-recyclable materials.

Machine learning model to predict solar energy output and evaluate battery storage strategies to reduce carbon emissions. Developed as part of the IE Sustainability Datathon in collaboration with Repsol. The model uses historical weather and solar generation data to forecast energy production and optimize storage allocation.