Alexander Kosak
Software Development Engineer at Satellite Innovations
Belarus
Summary
🤩
Rockstar🎓
Top SchoolAlexander Kosak is a software development engineer with 10 years' experience building production-grade systems across gaming, mapping, and satellite domains, currently engineering at Satellite Innovations. He brings deep practical expertise in computer vision and neural networks, evidenced by meaningful contributions to the widely used albumentations image-augmentation library—fixing color-space bugs, adding robust augmentations, and improving target-dependent transforms. His background spans embedded and mobile development up through backend and platform work at Wargaming and Mapbox, giving him a rare full-stack perspective on performance-sensitive systems. Fluent in both applied ML tooling and low-level implementation details, he pairs a Master’s in systems design with hands-on engineering that consistently turns research-grade techniques into reliable production features.
10 years of coding experience
13 years of employment as a software developer
Магистр, Системное проектирование, Магистр, Системное проектирование at Belarusian State UniversityБакалавр, Разработка программного обеспечения, Бакалавр, Разработка программного обеспечения at Belarusian State University of Informatics and Radioelectronics
Github Skills (10)
computer-vision10
opencv10
image-augmentation10
python10
image-processing9
numpy9
machine-learning9
deep-learning8
deeplearning-ai8
object-detection7
Programming languages (6)
Github contributions (5)
albumentations-team/albumentations
Jun 2018 - Aug 2022
Fast and flexible image augmentation library. Paper about the library: https://www.mdpi.com/2078-2489/11/2/125
Role in this project:
ML Engineer
ML EngineerContributions:116 commits, 98 PRs, 143 pushes in 4 years 2 months
Contributions summary:Alexander's commits primarily focused on improving the `albumentations` library by fixing and refactoring image augmentation techniques. The changes include corrections to HSV color space transformations, updates to random crop and resize augmentations, and adding a mechanism for making augmentations dependent on other targets (e.g., bboxes). Furthermore, they added new augmentation techniques, such as RandomSizedCrop, and improved the implementation and testing of existing augmentations.
fast-augmentationspythonalbumentationsdetectiondeep-learning
selimsef/dsb2018_topcoders
Apr 2018 - Sep 2018
DSB2018 [ods.ai] topcoders
Contributions:19 commits, 1 PR, 1 push in 4 months
machine-learningods
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.