Michael Wang
Founding AI Engineer at CarbonAI
Chicago, Illinois, United States
Summary
🤩
Rockstar🎓
Top SchoolMichael Wang is a Founding AI Engineer with a decade of experience turning messy data into production ML systems that drive revenue, from HFT trading desks to enterprise SaaS and early-stage startups. He specializes in 0-to-1 product development and has led teams to deliver high-accuracy, low-latency solutions—most notably production transformer use-cases and a transaction-to-brand entity resolution model that materially increased linked revenue. Technically hands-on, he’s contributed to open-source PyTorch time-series tooling with improved uncertainty estimation and interpretability (Plotly/Wandb visualizations and SHAP heatmaps). Based in Chicago, he blends algorithmic rigor (MS in Data Science) with product-minded engineering and a track record of turning direct reports into high-impact contributors. He’s drawn to data problems with real-world impact, particularly those that make systems smarter and everyday life better.
10 years of coding experience
10 years of employment as a software developer
University of California, San DiegoNanodegree Machine Learning, Nanodegree Machine Learning at UdacityMasters of Computer Science Data Science, Masters of Computer Science Data Science at University of Illinois Urbana-Champaign
English, Chinese, Japanese
Github Skills (14)
forecasting10
pytorch10
machine-learning10
forecast10
time-series10
deep-learning10
anomaly-detection10
python10
testing10
wandb9
plotly9
shap9
lstm8
transformers8
Programming languages (2)
Github contributions (5)
AIStream-Peelout/flow-forecast
May 2020 - Jul 2020
Deep learning PyTorch library for time series forecasting, classification, and anomaly detection (originally for flood forecasting).
Role in this project:
ML Engineer
ML EngineerContributions:38 commits, 4 PRs, 21 pushes in 1 month
Contributions summary:Michael primarily contributed to the evaluation and enhancement of a PyTorch-based time series forecasting library. Their work involved refactoring prediction functions, generating prediction samples for dropout layers, and adding comprehensive unit tests for evaluating model performance. They also made improvements to the training loop by integrating early stopping and incorporated plotting functionalities using Plotly and Wandb to visualize model outputs, including confidence intervals and heatmaps of Shap values, to improve model interpretability and provide actionable insights.
forecastingtime-series-analysistime-seriesclassificationautoencoder
michaelwang1994/Main
Mar 2016 - Mar 2022
Contributions:133 pushes, 1 branch in 6 years 1 month
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.