Justin Johnson
Cofounder at World Labs
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolWan-Yen Lo is a Senior Research Engineering Manager at Facebook AI Research in Mountain View with a PhD in Computer Science from UC San Diego and 11 years of industry experience. She leads a 20+ person West Coast Computer Vision team that bridges foundational research and product impact, contributing to high-profile work and frameworks such as Mask R-CNN, MAE, Detectron2 and PyTorch3D. Her background spans applied ML in safety-critical self-driving teams at Waymo and Google as well as search quality at Google Zürich, giving her deep experience shipping research into production. A hands-on engineer, she has made substantial open-source contributions—from CUDA kernels and nearest-neighbor optimizations in PyTorch3D to refinements of neural style and RNN toolkits—demonstrating comfort from low-level performance tuning to model design. Early research in stereoscopic imaging and motion planning plus tooling work like an OSX ulogme port underline a long-standing blend of systems pragmatism and vision research.
11 years of coding experience
7 years of employment as a software developer
Illinois Mathematics and Science AcademyBachelor of Science (BS), Mathematics and Computer Science, Bachelor of Science (BS), Mathematics and Computer Science at California Institute of TechnologyDoctor of Philosophy (PhD), Computer Science, Doctor of Philosophy (PhD), Computer Science at Stanford University
Github Skills (60)
pytorch10
content-management-system10
fasterrcnn10
benchmark10
maintenance10
c-language10
python10
scripting10
image-processing10
testing10
machine-learning10
benchmarking10
rnn-model10
n10
appkit10
Programming languages (7)
Github contributions (5)
jcjohnson/fast-neural-style
Oct 2016 - Jul 2017
Feedforward style transfer
Role in this project:
ML Engineer
ML EngineerContributions:52 commits, 11 PRs, 21 pushes in 9 months
Contributions summary:Justin primarily focused on modifying and improving the existing neural style transfer implementation. Their contributions include adjusting model configurations, updating loss function parameters, and adding scripts to assist with model evaluation and training processes. Furthermore, the user made changes to default settings, and checkpoint handling, indicating an active role in refining the training pipeline and model behavior. These modifications directly impacted the core functionality and efficiency of the style transfer process.
feedforwardstyle-transfergenerative-model
jcjohnson/densecap
Dec 2015 - Jun 2017
Dense image captioning in Torch
Role in this project:Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:126 commits, 5 PRs, 15 pushes in 1 year 6 months
Contributions summary:Justin made several contributions to the `densecap` repository, focusing on model and script refactoring. They modified and updated the `DenseCapModel.lua` and `LanguageModel.lua` files, likely related to the core deep learning model. Additionally, the user worked on the image processing and feature extraction scripts (`preprocess.py` and `extract_features.lua`), and integrated these with the test time and webcam demo scripts.
image-captioningtorchocrcomputer-visioncaptioning
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.