Ross Wightman
Angel Investor at Self-Employed
Vancouver, British Columbia, Canada
Summary
🤩
Rockstar🎓
Top SchoolRoss Wightman is an engineering leader, entrepreneur, and angel investor with 20+ years building high-performance imaging, robotics, and AI systems, and a decade of focused experience in machine learning and computer vision. He combines hands-on model engineering—contributing to prominent open-source projects like Hugging Face’s pytorch-image-models and OpenCLIP—with deep systems expertise in low-latency IO, distributed consensus, and FPGA/firmware integration. As a serial freelancer, founder and former director at Avigilon, he has shipped camera-to-cloud products and led rapidly growing cross-functional teams. Ross often backs and advises B2B hardware+software startups, preferring novel AI applications, and continues to prototype robotics and vision ideas seeking co-founder/CTO roles. An interesting detail: he ranks in the top 0.2% on Kaggle and has driven contributions that improved torchscript compatibility and distilled ViT model support used broadly in the ML community.
10 years of coding experience
17 years of employment as a software developer
BASc (Honours), Computer Engineering, BASc (Honours), Computer Engineering at Simon Fraser University
Stackoverflow
747reputation
105kreached
4answers
0questions
tensorflow
top-5%
machine-learning
top-5%
deep-learning
top-5%
keras
top-5%
Github Skills (26)
pytorch10
caffe10
integrate10
python10
machine-learning10
onnx10
integrations10
refactor10
deep-learning10
mobilenet10
trainings10
object-detection10
computer-vision10
refactoring10
efficientnet10
Programming languages (9)
Github contributions (5)
mlfoundations/open_clip
Jan 2022 - Jan 2023
An open source implementation of CLIP.
Role in this project:
ML Engineer
ML EngineerContributions:2 releases, 28 reviews, 143 commits in 1 year
Contributions summary:Ross made significant changes to the codebase, primarily focused on refactoring and improving the implementation of CLIP, an open-source implementation of a Contrastive Language-Image Pre-training model. Their contributions included replacing CUDA usage, making the main script single-process, integrating timm-based vision configurations, cleaning up loss functions, and adding support for distributed training using Horovod and potentially SLURM. They were also involved in refactoring model code and adding new model types.
pytorchlanguage-modelzero-shot-classificationdeep-learningclip
rwightman/gen-efficientnet-pytorch
May 2019 - Jul 2021
Pretrained EfficientNet, EfficientNet-Lite, MixNet, MobileNetV3 / V2, MNASNet A1 and B1, FBNet, Single-Path NAS
Role in this project:ML Engineer
ML EngineerContributions:104 commits, 13 PRs, 80 pushes in 2 years 2 months
Contributions summary:Ross's contributions primarily revolve around implementing and integrating various deep learning models, particularly focusing on architectures related to EfficientNet, MobileNetV3, and MixNet. They added code for the inclusion of pretrained weights for FBNet-C, MobileNet-V3, and EfficientNet models. The user also developed scripts for ONNX export and Caffe2 validation, demonstrating a focus on model deployment and compatibility across different frameworks.
pytorchmobilenetcaffe2efficientnetdeep-learning
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.