Vraj Patel

Automated analysis of 38 years of estuary data (10K+ records) across 7 sites, building scalable workflows that cut manual effort by 70%.

Digitized and standardized 8 years of legacy data (1987–1994) using OCR, consolidating into a centralized SQL repository for unified access.

Developed PowerQuery ETL pipelines, reducing report turnaround from 3 weeks to 10 minutes (95% faster).

Created VBA-based 1-click reporting, automating data distribution to estuary-specific dashboards.

Built and deployed 4 interactive dashboards (trend analysis, anomaly detection, inter-estuary comparisons).

Implemented automated exception reporting, improving accuracy and visibility for operational planning.

Produced detailed SOPs and training guides, ensuring 100% knowledge transfer for future interns and staff.

Developed an object detection pipeline using YOLOv2 in MATLAB Deep Learning Toolbox, achieving 82.4% mAP on coral reef imagery.

Preprocessed and augmented a dataset of 1,000+ underwater images using normalization, adaptive histogram equalization (AHE), and cropping to enhance robustness.

Utilized MATLAB’s Image Labeler, DataStore, and GPU modules for efficient model training and validation.

Adapted the YOLOv2 architecture for real-time 720p video tracking, enabling continuous marine species detection.

Validated performance on multiple reef videos with consistent detection of coral species, clownfish, and sea turtles.

Open-sourced the full project on GitHub: Object-Detection-using-YOLOv2.

Collaborating with IT team to resolve technical issues, providing support and enhancing user satisfaction. Over 1+ Years of customer service experience.