Teddy Koker
PHD Student at MIT EECS
Cambridge, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolTeddy Koker is a machine learning researcher and PhD student at MIT with 11 years of experience building and productionizing ML models and tooling. He has a strong track record at MIT Lincoln Laboratory and Lightning AI, contributing to core open-source projects like PyTorch Lightning and TorchMetrics and implementing models such as Image GPT and Faster R-CNN. Teddy combines research rigor with engineering pragmatism—improving TPU dataloaders, gradient clipping tests, and metric implementations while adding practical demos and datamodules for reproducible workflows. Based in Cambridge, he blends academic depth in EECS with hands-on engineering from industry internships to staff roles, often focusing on scalable training and evaluation infrastructure.
11 years of coding experience
4 years of employment as a software developer
Doctor of Philosophy - PhD, Electrical Engineering and Computer Science, Doctor of Philosophy - PhD, Electrical Engineering and Computer Science at Massachusetts Institute of TechnologyBachelor of Science - BS, Computer Science, Bachelor of Science - BS, Computer Science at Worcester Polytechnic Institute
Stackoverflow
964reputation
341kreached
14answers
0questions
timer
top-5%
html
top-5%
Github Skills (27)
pytorch10
python10
image-processing10
data-science10
image-classification10
testing10
machine-learning10
pytorch-lightning10
deeplearning-ai10
deep-learning10
ai10
computer-vision10
metric10
classification9
timer9
Programming languages (8)
Github contributions (5)
Lightning-Universe/lightning-flash
Aug 2020 - Feb 2021
Your PyTorch AI Factory - Flash enables you to easily configure and run complex AI recipes for over 15 tasks across 7 data domains
Role in this project:
ML Engineer
ML EngineerContributions:35 reviews, 47 commits, 15 PRs in 5 months
Contributions summary:Teddy made several commits focused on refactoring and reorganizing the Flash framework, including the test structure, and import statements. They also removed an argument from the flash class. Furthermore, they added a base demo notebook that demonstrates the usage of the flash library. They added a Kaggle image classification demo and made changes to the image classification datamodule.
torch-geometricicevisionconfigurepytorch-lightningdomains
Lightning-AI/pytorch-lightning
Mar 2020 - Feb 2021
Pretrain, finetune ANY AI model of ANY size on multiple GPUs, TPUs with zero code changes.
Role in this project:ML Engineer
ML EngineerContributions:128 reviews, 18 commits, 32 PRs in 10 months
Contributions summary:Teddy contributed significantly to the PyTorch Lightning framework, focusing on improving its machine learning capabilities. Their work included fixing dataloaders for TPU examples, adding tests for gradient clipping with native AMP, and implementing an Explained Variance metric. The user also refactored and improved existing metrics, enhanced documentation, and added new features for model training and evaluation. This involved changes to core trainer functionality and metric integration, alongside contributions for efficient model training.
pythonheadachespytorch-modelsdata-sciencehandling
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.