Andrey Malevich
Member Of Technical Staff at OpenAI
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolAndrey Malevich is a seasoned machine learning and infrastructure leader with nine years building recommendation and graph learning systems at scale, most recently as a Member of Technical Staff at OpenAI after directing Graph Learning for Ads at Meta. He has driven org-level strategy and productionization of sequence and graph models that power Meta-scale personalization, co-authoring large-scale ML systems that materially improved ad prediction accuracy and revenue. Hands-on across backend, ML and platform work, Andrey contributed to Caffe2 core operators—improving robustness and performance—which underscores his ability to bridge research, open-source frameworks, and production engineering. Based in San Francisco, he combines deep CS training from Belarusian State University and Yandex School of Data Analysis with a track record of turning state-of-the-art models into highly optimized, business-critical systems.
9 years of coding experience
11 years of employment as a software developer
Computer Science, Computer Science at Belarusian State UniversityYandex School of Data AnalysisHigh School Mathematics, High School Mathematics at Belarussian State University Lyceum
Russian, English
Github Skills (13)
operator10
machine-learning10
caffe10
deep-learning10
python10
ai10
back-end-development10
cprogramming-language9
computer-engineering9
c-language9
performance-optimization8
pytorch3
tensorflow3
Programming languages (4)
Github contributions (5)
facebookarchive/caffe2
Nov 2016 - Feb 2018
Caffe2 is a lightweight, modular, and scalable deep learning framework.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:82 commits in 1 year 2 months
Contributions summary:Andrey primarily worked on fixing operator implementations within the Caffe2 deep learning framework, addressing issues such as handling empty batches and ensuring correct functionality of various operators like convolutions, pooling, and LRN. They also made improvements to serialization speed and addressed bugs related to parameter handling and shape inference, focusing on code related to the core framework functionality. Furthermore, the user contributed to enhancing the framework's capabilities by adding support for more data types and improving the handling of specific layer types, such as SparseToDense and Recurrent Networks.
pytorchscalablecaffe2deep-learningml
kennyhorror/pytorch
Oct 2018 - Sep 2019
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:9 pushes, 7 branches in 11 months
pythongpu-accelerationdeep-learninggpuacceleration
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.