Michael Diskin
Deputy Head Of LLM Section at Higher School of Economics
Yerevan, Armenia
Summary
🤩
Rockstar🎓
Top SchoolMichael Diskin is a seasoned AI and deep learning leader with 11 years of experience, currently serving as Deputy Head of the LLM Section at Wildberries where he manages a 26-person team and oversees scalable text-processing infrastructure and cross-team LLM strategy. He blends hands-on research and engineering—having contributed to distributed deep learning at Yandex and to notable open-source projects like the decentralized training framework hivemind and a Practical RL course—while co-authoring papers at top ML conferences. Michael’s background spans academia and industry, including tutoring at Yandex School of Data Analysis and pursuing a PhD in Computer Science, which informs his emphasis on mentorship and reproducible R&D. Pragmatic about production constraints, he’s addressed real-world issues such as CI modernization, dependency management, and operational limits in large-scale ML systems.
11 years of coding experience
3 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Higher School of EconomicsMaster's degree, Artificial Intelligence, Master's degree, Artificial Intelligence at Yandex School of Data Analysis
Russian, English, French
Github Skills (23)
asynchronous10
pytorch10
python10
async10
distributed-systems10
reinforcement-learning10
cicd10
deep-learning10
deep-reinforcement-learning10
jupyter-notebook10
dht10
devops10
asyncio9
deeplearning-ai9
docker8
Programming languages (4)
Github contributions (5)
learning-at-home/hivemind
Feb 2021 - Feb 2022
Decentralized deep learning in PyTorch. Built to train models on thousands of volunteers across the world.
Role in this project:
Back-end & DevOps Engineer
Back-end & DevOps EngineerContributions:61 reviews, 81 commits, 34 PRs in 11 months
Contributions summary:Michael made significant contributions to the `hivemind` project, primarily focused on improving its infrastructure and core functionality. Their work involved updating build processes and dependencies in `setup.py`, adding testing for DHT functionality, and optimizing metrics collection. Further commits included enhancing the build process, and switching CI to GitHub Actions, and adding support for auxiliary peers. The user also addressed the "too many open files" issue.
pytorchhiveminddhtasynciovolunteers
yandexdataschool/Practical_RL
May 2019 - Mar 2022
A course in reinforcement learning in the wild
Role in this project:ML Engineer
ML EngineerContributions:3 reviews, 37 commits, 60 PRs in 2 years 10 months
Contributions summary:Michael contributed to backporting modifications from a Coursera version of a reinforcement learning course. These changes involved updating code for various weeks of the course, including introductory concepts, value-based methods, and policy-based methods. Specifically, the commits include updates to notebooks and Python scripts related to the course content and examples.
pytorchgit-coursedeep-learningreinforcement-learningwild
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.