Alexander Rakhlin
Senior Researcher at Neuromation
Contact Info
Other
Summary
👤
Senior🎓
Top SchoolAlexander Rakhlin is a Senior Researcher and machine learning specialist with 11+ years applying deep learning and classical ML to medicine, science and business problems. A Kaggle Master (ranked ~129 of 66,000) and winner of CERN’s “Physics Prize: HEP meets ML,” he has repeatedly placed highly in diverse competitions from medical imaging to geospatial and forecasting tasks. At Neuromation and as an independent consultant he has delivered production-grade models and research—examples include winning and improved MICCAI robotic instrument segmentation code and CNNs for sentence classification in Keras. His toolkit spans Python, scikit-learn, Keras/Theano/Caffe and OpenCV, and he pairs strong math/physics training (MS in CS, background in physics and mathematics) with practical engineering, including an automated derivatives trading system he architected and deployed. Colleagues value his ability to translate competitive research into robust, explainable solutions for real-world data.
10 years of coding experience
Mathematics, Mathematics at School of Mathematics at Lomonosov Moscow State UniversityM.Sc. in Computer Science and Microelectronic Devices, M.Sc. in Computer Science and Microelectronic Devices at National Research University of Electronic Technology (MIET)Physics, Physics at School of Physics at St.Petersburg State University
English, Russian, Esperanto
Github Skills (17)
pytorch10
opencv10
python10
machine-learning10
mask-rcnn10
keras10
image-segmentation10
deeplearning-ai10
deep-learning10
segmentation10
medical-imaging10
faster-rcnn10
computer-vision10
nlp10
word2vec9
Programming languages (8)
Github contributions (5)
alexander-rakhlin/CNN-for-Sentence-Classification-in-Keras
Mar 2016 - Jul 2017
Convolutional Neural Networks for Sentence Classification in Keras
Role in this project:
ML Engineer
ML EngineerContributions:14 commits, 2 PRs, 23 pushes in 1 year 4 months
Contributions summary:Alexander focused on developing and refining a convolutional neural network (CNN) for sentence classification using Keras. They added prediction functionality to the model, allowing it to classify new sentences based on the trained model. The user also made significant changes, including switching to a larger IMDB corpus, redesigning the code, and upgrading to Keras 2.0. Furthermore, they addressed initialization issues and added functionality to load data from a local directory.
sentencesentence-classificationkeras-modelsdeep-learningconvolutional
ternaus/robot-surgery-segmentation
Feb 2018 - Mar 2018
Wining solution and its improvement for MICCAI 2017 Robotic Instrument Segmentation Sub-Challenge
Role in this project:ML Engineer
ML EngineerContributions:26 commits, 3 pushes, 1 comment in 17 days
Contributions summary:Alexander primarily focused on modifying and extending image processing code for the purpose of robotic instrument segmentation. The commits introduce functionality to overlay segmentation masks onto the original images, using different modes such as "BINARY," "PARTS," and "TYPE". They implemented code for image resizing and utilized OpenCV (cv2) to manipulate images, demonstrating a focus on computer vision techniques relevant to the project's goals. The user also contributed to generating image grids for visualization, effectively demonstrating expertise in model output analysis within a medical image domain.
improvementsegmentationcomputer-visionmiccaideep-learning
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.