Michael Wyatt
Project Management Lead at Strum Platform
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolMichael Wyatt is a seasoned project management leader with over 20 years of experience delivering complex programs across finance, healthcare, manufacturing, and cloud technology. Currently leading project management at Strum Platform after roles at Microsoft and Global Credit Union, he blends PMI-ACP and Scrum practice with hands-on Azure administration to drive on-time, on-budget outcomes. He has led cross-functional teams from startup founders to enterprise stakeholders and built and sold custom software through his own company, demonstrating both delivery rigor and entrepreneurial instincts. An active contributor to ML/MLops open-source—fixing inference, CUDA/PyTorch compatibility, and deployment issues in DeepSpeed and DeepSpeed-MII—he brings uncommon technical fluency for a PM. Known for improving ROI and operational quality, he pairs lean workflow and scrum leadership with an eye for production-grade ML deployment. Based in Seattle and educated at Michigan, he thrives at the intersection of cloud infrastructure, software delivery, and machine learning operations.
10 years of coding experience
26 years of employment as a software developer
Michigan
English
Github Skills (16)
pytorch10
machine-learning10
deeplearning-ai10
deepspeed10
deep-learning10
inference10
pytest10
python10
testing10
dockers9
transformers9
cuda9
docker9
pydantic8
gpu7
Programming languages (3)
Github contributions (5)
deepspeedai/DeepSpeed-MII
May 2022 - Jan 2023
MII makes low-latency and high-throughput inference possible, powered by DeepSpeed.
Role in this project:
MLOps Engineer
MLOps EngineerContributions:7 releases, 165 reviews, 44 commits in 8 months
Contributions summary:Michael primarily contributed to the deployment and configuration aspects of the DeepSpeed-MII project. Their work involved integrating Pydantic for validating configurations, establishing unit tests for local deployment scenarios, and fixing bugs related to model deployment, especially for models utilizing FP16 and Bloom. They also updated the repository to support new Transformers versions and the new cache design, and also enhanced the existing AML deployment scripts and configurations.
pytorchdeepspeeddeep-learninginferencelatency
deepspeedai/DeepSpeed
Mar 2022 - Jan 2023
DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
Role in this project:ML Engineer
ML EngineerContributions:12 releases, 438 reviews, 103 commits in 10 months
Contributions summary:Michael primarily contributed to the testing and improvement of the DeepSpeed library, with a focus on inference tasks. They added unit tests for various model families and inference tasks, including HF models. The user fixed import errors and addressed compatibility issues, particularly related to CUDA and PyTorch versions. Additionally, they refactored configurations for ZeRO and made improvements to model profiling, indicating a strong understanding of DeepSpeed's core functionality and optimization strategies.
billion-parametersfinetuningtrainingmixture-of-expertszero
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.