At Trimble Inc., I developed and released TrimbleDL using C# Xamarin and .NET MAUI for cross-platform mobile delivery. I also developed a GPT-4 solution that generates XSLT code from customer requirements for Trimble Access report generation.
I built a cloud video-processing workflow with React, AWS S3, Lambda, and EC2, automating Python processing and uploading annotated videos for frontend display. I also experimented with U-Net, ResNet, and YOLO models for road pothole detection.
My research at Anna University focused on retinal vessel segmentation with supervised and unsupervised machine learning, using MATLAB and Python. Recent GitHub projects apply LLM workflows and computer vision to PDF question processing and semiconductor defect analysis.

