Machine Learning Engineer

Wei Wang

Madison, WI 608-***-**** ********@****.*** https://github.com/wwang487

https://www.linkedin.com/in/wei-wang-199596350/

Education

University of Wisconsin–Madison, Ph.D., Civil and Environmental Engineering, Expected August 2025 GPA:3.89 University of Wisconsin–Madison, M.S., Computer Science, 2021 GPA:3.90 Peking University, B.S., Environmental Science, and B.A., Economics, 2017 GPA 3.72 Technical Skills

- Programming Languages: Python, Java, Matlab, C++, JavaScript, R

- Frameworks & Libraries: PyTorch, TensorFlow, OpenCV, Keras, Scikit-learn, MySQL, Node.js

- Tools & Systems: Git, Linux, Spark, Docker, AWS, Azure, Remote Sensing, Structure-from-Motion Research Experience

Coastal Feature Analysis Toolkit (CFAT) Deep Learning Engineer 2022-2025

• Built a large-scale shoreline segmentation dataset by labeling 20,689 aerial images.

• Trained a DeepLabV3+ model for water segmentation, achieving 98.9% pixel accuracy and 92.3% mIoU.

• Designed a custom loss function using mean neighboring shoreline covering ratio (mNSCR), boosting model performance by over 1%.

• Developed a C++ module (GDAL/OpenCV) for shoreline extraction and movement rate calculation, integrated with Python via ctypes for 80% faster processing than ArcMap/QGIS.

• Enabled Spark-based deployment of CFAT to support distributed shoreline analysis over large aerial datasets, such as full-basin processing for the Great Lakes.

• Fine-tuned LLaMA3 and LLaVA on hand-labeled coastal imagery and descriptions, achieving 86.4% accuracy in identifying failures of coastal structures.

Dangerous Current Detection & Warning System Deep Learning Engineer 2019–2023

• Deployed LOCKS, a real-time coastal monitoring system capturing images every 10 seconds via 2 Mbps Ethernet-connected cameras.

• Labeled 8,013 ortho-images to build a large-scale dangerous current detection dataset.

• Trained and refined a Cascade R-CNN model, achieving 96.4% precision and 93.4% recall.

• Designed post-processing algorithms that boosted detection precision by 11.3%.

• Integrated the model into LOCKS for sub-2s real-time inference and automated wireless public alerting.

• Contributed to four successful drowning rescues in Ozaukee County, WI through early detection and response. Product Development Experience

Jicunbao –WeChat Mini Program for Shared Storage Tech Lead 2016–2017

• Developed a WeChat Mini Program for shared public storage access using JavaScript and WXML/WXSS, with QR code login and WeChat Pay integration.

• Built backend services with Node.js and MySQL and deployed on WeChat Cloud.

• Led technical development and campus marketing, reaching 450 peak daily users.

• Raised ¥800,000 RMB (~$100,000 USD) in funding to scale the platform. Teaching Experience

- Problem solving using computer tools (Python), UW Madison, 2022-2025

- Problem solving using computer tools (Matlab), UW Madison, 2018-2022 Honors & Awards

- Becker's Graduate Student Conference Travel Grant (2024, 2022)

- Outstanding Student Academic Scholarship, Peking University (Ranked 1st)

Contact this candidate