Alexander Matyasko
Research Fellow at Singapore Institute of Technology
Singapore, Singapore
Summary
👤
Senior🎓
Top SchoolAlexander Matyasko is a research-focused machine learning engineer with 11 years of experience building robust, GPU-accelerated models and production-ready tooling across academia and industry. Currently a Research Fellow in Singapore, he has worked on adversarial robustness, federated learning defenses, and physics-informed deep learning while contributing to landmark open-source projects like Theano and the CleverHans adversarial library. His background blends low-level C/CUDA work for GPU pooling kernels with high-level model engineering—adding virtual adversarial training to major ML toolkits and integrating modern LLMs (including Google’s Gemini) into self-hostable systems. He has repeatedly shipped real-time vision and point-cloud solutions for applied problems (fire/smoke detection, 3D tracking) and adapted large-scale dataset converters and tests for reproducible pipelines. Comfortable across back-end, full-stack, and research roles, he pairs PhD-level rigor with pragmatic engineering that tightens the loop from prototypes to deployable systems. An understated strength is his knack for improving developer UX in research codebases, from robust unit tests to editor integrations that accelerate productivity.
11 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (PhD), Computer Science, Doctor of Philosophy (PhD), Computer Science at Nanyang Technological UniversityDiploma of Engineer, Applied Mathematics and Computer Science, Diploma of Engineer, Applied Mathematics and Computer Science at Belarusian State University of Informatics and Radioelectronics
English, Russian, Russian
Github Skills (42)
adversarial-attacks10
data-conversion10
data-pipelines10
python10
gpu-programming10
machine-learning10
configuration-management10
c1110
numpy10
security10
c1710
ml10
tensorflow10
emacs10
configuration-file10
Programming languages (7)
Github contributions (5)
Theano/Theano
Oct 2016 - Jul 2017
Theano was a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. It is being continued as PyTensor: www.github.com/pymc-devs/pytensor
Role in this project:
Back-end Developer
Back-end DeveloperContributions:96 commits, 10 PRs, 126 comments in 9 months
Contributions summary:Alexander primarily contributed to the Theano project by implementing and improving GPU-based pooling operations, focusing on max and average pooling. Their work involved writing and modifying C code for the `gpuarray` backend, including kernel implementations and support code for these pooling methods. They also added and refined testing for the implemented GPU operations.
mathematical-expressionsmultidimensional-arrayspymcpythontheano
cleverhans-lab/cleverhans
Apr 2017 - Jun 2017
An adversarial example library for constructing attacks, building defenses, and benchmarking both
Role in this project:ML Engineer
ML EngineerContributions:10 commits, 2 PRs, 15 comments in 1 month
Contributions summary:Alexander primarily contributed to implementing and testing virtual adversarial training (VAT) methods within the CleverHans library, which focuses on adversarial example generation. They added both Theano and TensorFlow implementations of VAT, incorporating changes to existing attack methods to support backend flexibility. Furthermore, the user integrated unit tests for the newly developed utility functions and VAT methods, ensuring the robustness of these adversarial training implementations. They also refined the code by correcting constant type definitions for better accuracy.
benchmarkingmachine-learningsecurity
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.