Simon Niklaus
Staff Research Scientist at Google DeepMind
Vancouver, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolSimon Niklaus is a Staff Research Scientist with 11 years of experience at the intersection of computer vision and machine learning, now working on GenAI data at Google DeepMind after leading foundational research and product launches at Adobe Firefly. He has a strong track record shipping production features—image denoising, lens blur, sky segmentation, optical flow, and creative mockup tools—and his open-source work includes PyTorch implementations of PWC-Net, separable-convolution SLOMO, and a 3D Ken Burns effect that highlight deep expertise in video interpolation and efficient model/kernel design. A PhD candidate in computer science, Simon combines rigorous academic training with hands-on CUDA and PyTorch engineering, often optimizing kernels and cost-volume layers for real-world performance. Notably, he was one of five founders of Adobe Firefly’s data efforts and continues to publish practical, well-engineered repos that bridge research prototypes and production systems.
11 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at Portland State UniversityMaster of Science (M.Sc.) Information Systems, Master of Science (M.Sc.) Information Systems at Technical University of Applied Sciences Würzburg-Schweinfurt (THWS)
German, English
Github Skills (15)
cuda10
computer-vision10
pytorch10
deeplearning-ai10
deep-learning10
python10
neural-network9
machine-learning9
convolutional-neural-networks9
c-language8
faster-rcnn8
mask-rcnn8
video-processing8
image-processing8
cprogramming-language8
Programming languages (11)
Github contributions (5)
sniklaus/3d-ken-burns
Nov 2019 - Mar 2022
an implementation of 3D Ken Burns Effect from a Single Image using PyTorch
Role in this project:
ML Engineer
ML EngineerContributions:43 commits, 5 PRs, 44 pushes in 2 years 4 months
Contributions summary:Simon's contributions primarily involve modifications to the codebase related to 3D Ken Burns effect implementation using PyTorch. The commits demonstrate changes in various files including model definitions, kernel implementations, and benchmark scripts, suggesting the user is working on model architecture and performance optimizations. The user also modified the code to ensure that the project is compatible with pytorch version 1.2.0 and later.
cudapytorchpythondeep-learningcupy
sniklaus/sepconv-slomo
Sep 2017 - Nov 2022
an implementation of Video Frame Interpolation via Adaptive Separable Convolution using PyTorch
Role in this project:ML Engineer
ML EngineerContributions:58 commits, 4 PRs, 55 pushes in 5 years 2 months
Contributions summary:Simon primarily contributed to the implementation of a video frame interpolation system using adaptive separable convolution. Their work involved modifications to the core `SeparableConvolution` layer, including CUDA kernel integration and forward pass logic. The user also made changes to the main `run.py` script, suggesting involvement in model training and evaluation procedures.
pytorchframeadaptiveoptical-flowconvolution
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.