Matěj Račinský
Senior Researcher at Gen™
Prague, Prague, Czechia
Summary
👤
Senior🎓
Top SchoolMatěj Račinský is a Senior Researcher and machine learning engineer with 11 years of experience building and validating deep learning systems, from computer vision for industrial defect detection to time-series forecasting for energy applications. Based in Prague, he blends research rigor with hands-on engineering—contributing numeric-stability improvements and derivative rules to the widely used FluxML/Zygote.jl autodiff project while ensuring robustness through test automation in Flux.jl. His background spans academia and industry, including generative models and synthetic-data pipelines extracted from simulation environments, and he’s comfortable moving models from prototype to production. Known for meticulous testing and performance-minded implementations, he favors practical solutions that make data meaningful and reliable.
11 years of coding experience
6 years of employment as a software developer
Master's degree, Artificial Intelligence, 2., Master's degree, Artificial Intelligence, 2. at Faculty of Electrical Engineering, Czech Technical University in Prague
Stackoverflow
1,708reputation
102kreached
64answers
20questions
Github Skills (27)
testing10
machine-learning10
gradient10
automatic-differentiation10
flux10
neural-network10
fluxor10
julia10
data-science9
numerical9
artificial-neural-networks9
numerical-methods9
deeplearning-ai9
deep-learning9
numerics9
Programming languages (23)
Github contributions (5)
FluxML/Flux.jl
Oct 2020 - Oct 2021
Relax! Flux is the ML library that doesn't make you tensor
Role in this project:
QA Engineer / Test Automation Engineer
QA Engineer / Test Automation EngineerContributions:14 reviews, 37 commits, 5 PRs in 1 year
Contributions summary:Matěj primarily contributed to the testing aspects of the Flux.jl repository. Their commits involved writing and modifying tests to ensure the correctness of the library's functionality. They specifically addressed issues related to the `show` method and data loading, demonstrating a focus on verifying existing code and uncovering potential bugs. The user also introduced tests for new features like data loaders and ensured proper model training and expected outputs.
machine-learningtensorfluxneural-networksdeep-learning
FluxML/Zygote.jl
Jul 2020 - Oct 2020
21st century AD
Role in this project:ML Engineer
ML EngineerContributions:2 reviews, 17 commits, 3 PRs in 2 months
Contributions summary:Matěj primarily contributed to the `zygote.jl` repository by implementing and optimizing automatic differentiation functionalities related to neural network layers and activation functions. This included the addition of derivative rules for functions like SELU, Tanh, and ELU, alongside the introduction of test cases to validate the correctness and numerical stability of the implemented gradients. Furthermore, the user addressed specific issues and refined existing code, focusing on improvements related to numeric stability and performance.
control-flowautomatic-differentiationmachine-learningjulia-compilerjulia
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.