Arsenii Ashukha
Staff Research Scientist at Isomorphic Labs
London, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolArsenii Ashukha is a Senior Research Scientist in machine learning with 11 years of experience bridging academic rigor and industrial impact, currently building stealth AI models for drug design at Isomorphic Labs. He holds advanced degrees from MIPT, Yandex School, and a PhD-level trajectory at HSE, and has a track record of leading research teams and labs at Samsung and Yandex. His work spans generative models, uncertainty estimation, sparsification (notably Sparse Variational Dropout), and high-resolution image inpainting, with contributions to popular open-source projects such as LaMa. A seasoned educator, he co-authored ML course materials and taught deep learning at MIPT and HSE, making complex methods accessible to students and practitioners. Early projects include developing autoencoder models and latent-space analyses for MNIST and courseware, reflecting hands-on expertise in model design and evaluation. He combines deep theoretical insight with product-minded research, focusing on practical, scalable ML solutions for real-world scientific problems.
11 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science / Machine Learning, Doctor of Philosophy - PhD, Computer Science / Machine Learning at Higher School of EconomicsMaster’s Degree, Computer Science, Master’s Degree, Computer Science at Moscow Institute of Physics and Technology (State University) (MIPT)Master’s Degree, Machine Learning, Master’s Degree, Machine Learning at Yandex School of Data AnalysisBachelor’s Degree, Computer Science, Bachelor’s Degree, Computer Science at Bauman Moscow State Technical University
English, Russian
Github Skills (9)
lasagne10
machine-learning10
deep-learning10
python10
theano10
autoencoder10
data-analysis9
mnist9
tensorflow3
Programming languages (2)
Github contributions (5)
vkantor/MIPT_Data_Mining_In_Action_2016
Oct 2016 - Dec 2016
"Data Mining in Action Course", Moscow Institute of Physics and Technologies
Role in this project:
Data Scientist
Data ScientistContributions:54 commits, 1 PR, 49 pushes in 2 months
Contributions summary:Arsenii contributed to the development of an autoencoder model for MNIST handwritten digit recognition. The commits demonstrate the implementation and modification of the autoencoder, including its architecture, training, and evaluation, utilizing Python and the Lasagne library. The user further explored the latent space of the autoencoder through t-SNE projections and refactored the code to address a kaggle competition. The user's work focused on building and refining a deep learning model within a data mining context.
data-mining
senya-ashukha/bigram-anchor-words
Jan 2015 - Jun 2016
Contributions:135 commits, 1 PR, 153 pushes in 1 year 5 months
pythontool-kittopic-modelingmodelingmodeling-tool
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.