Alexander Kapitanov
Head Of ML R&D at Sberbank
Russia
Summary
🤩
Rockstar🎓
Top SchoolAlexander Kapitanov is Head of ML R&D with over a decade of experience building and scaling multidisciplinary AI teams and production-grade multimodal systems at Sberbank, where he led R&D for Russia’s first multimodal LLM and shipped products in vision, speech and sign-language recognition. He combines deep hands-on DSP and FPGA expertise—authoring open-source DSP materials and implementing high-performance FFTs and filter architectures—with modern ML engineering, enabling efficient edge deployments through distillation, pruning and automated data pipelines. A pragmatic leader and mentor, he has grown multiple teams from scratch, shepherded projects from research to product, and raised engineers into lead roles. His work produced unique datasets (HaGRID, Slovo) and delivered automated annotation and validation pipelines that dramatically reduced labeling costs while boosting confidence. Based in Russia, he blends academic rigor (MPEI background and publications) with practical systems know-how across hardware and large-scale ML.
11 years of coding experience
12 years of employment as a software developer
Master's degree, Radio, Television, and Digital Communication, Highest, Master's degree, Radio, Television, and Digital Communication, Highest at Moscow Power Engineering Institute (Technical University)
English, Russian
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (11)
digital-signal-processing10
convolution10
scipy10
dft10
fft10
python10
resampling10
signal-processing10
matplotlib10
numpy10
window-functions9
Programming languages (4)
Github contributions (5)
hukenovs/dsp-theory
Jul 2019 - Oct 2022
Theory of digital signal processing (DSP): signals, filtration (IIR, FIR, CIC, MAF), transforms (FFT, DFT, Hilbert, Z-transform) etc.
Role in this project:
Data Scientist
Data ScientistContributions:4 reviews, 67 commits, 14 PRs in 3 years 3 months
Contributions summary:Alexander contributed significantly to the development of digital signal processing theory, including lectures on signals, spectrum analysis, convolution, and resampling techniques. They appear to be focused on the creation and presentation of educational content, implementing various examples and visualizations using Python libraries. The user has worked on the development of polyphase FFTs and the use of window functions in signal processing.
signalpythonconvolutionnumbascipy
hukenovs/blackman_harris_win
Sep 2018 - Aug 2020
Contributions:72 commits, 63 pushes, 1 branch in 1 year 11 months
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.