Alexey Kruglov
Warsaw, Masovian Voivodeship, Poland
Summary
🤩
Rockstar🎓
Top SchoolAlexey Kruglov is an engineer with 10+ years of experience spanning deep learning, numerical simulation and theoretical physics, combining MSc-level physics rigor with production-grade software skills in Python, PyTorch and some C++. He designs and ships DL systems across computer vision (tracking, detection, few-shot) and speech recognition (including NLP spelling correction and CTC+LM decoding), and his work at OpenCV.ai and Intel produced measurable gains—e.g., a 1.5× improvement in tracking quality and ~2× pipeline speedup—and contributed speech recognition fixes and demos to Intel’s Open Model Zoo improving WER/CER. He has driven research-to-product flows (papers, patents, cross-team coordination) and has hands-on R&D experience in numerical simulation for industrial processes. A former competitive programmer and multiple prize-winner in math and physics competitions, he brings strong analytical instincts and is ready to grow into new fields or roles.
10 years of coding experience
9 years of employment as a software developer
with distinction (5.000 of 5.0), with distinction (5.000 of 5.0) at State University of Nizhni Novgorod named after N.I. Lobachevsky (UNN)
English, Russian, Polish
Stackoverflow
61reputation
2kreached
1answer
0questions
Github Skills (14)
machine-learning10
speech-recognition10
deeplearning10
inference10
deep-learning10
modello10
python10
openvino10
tensorflow9
cnn-model9
tensorflow29
c-programming-language7
c-language7
cuda6
Programming languages (4)
Github contributions (5)
openvinotoolkit/open_model_zoo
Aug 2020 - Jun 2021
Pre-trained Deep Learning models and demos (high quality and extremely fast)
Role in this project:
ML Engineer
ML EngineerContributions:96 reviews, 28 commits, 11 PRs in 9 months
Contributions summary:Alexey primarily contributed to the accuracy and functionality of speech recognition metrics within the Open Model Zoo repository, focusing on metrics like Word Error Rate (WER) and Character Error Rate (CER). Their work involved fixing and improving the implementations of these metrics, renaming and adding functionalities. The user also added a speech recognition demo and integrated adapters for CTC decoding with a language model, and integrated a public model. Further contributions involved fixing pylint warnings and adapting the project for Windows.
modelonnx-modelsdeep-learning-modelsmodel-zoopytorch-models
AlexeyKruglov/open_model_zoo
Aug 2020 - May 2021
Pre-trained Deep Learning models and samples (high quality and extremely fast)
Contributions:47 pushes, 14 branches in 9 months
pytorchcaffe2deep-learningunet-image-segmentationpre-trained
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.