At IGS Energy, I built Gemini API bill parsers in Python that automate 93% of data extraction from PDFs and scanned images. I also added automated evaluation in Azure DevOps CI pipelines to catch regressions before deployment.
I resolved more than 20 production issues by tracing parser failures to their causes and adding retry, fallback, and validation logic. I also set up logging, monitoring, and alerting, and delivered an asset hierarchy feature using a MySQL nested set model, REST API endpoints, and React workflows.
At The Ohio State University, I trained CNN-LSTM models in PyTorch for ECG classification, ranking in the Top 10 globally in the George B. Moody PhysioNet 2025 Challenge. I built data pipelines to process 366,000 biomedical records and tracked model experiments using scikit-learn cross-validation.
I also built the computer vision occupancy detection pipeline, Node.js Express API, and Selenium tests for OSU Smart Parking, which won 2nd place in the Honda Challenge at HackOHI/O 2025. As President of Collaborative Software Development Club, I lead four product teams and review members’ pull requests.
