Sergey Kolesnikov

Sergey Kolesnikov

R&D Lead

Tinkoff

Catalyst

Tl;dr

I am an R&D Lead at Tinkoff bank, working on various machine learning applications. I focus on the intersection of machine learning and product development to create novel solutions that can improve customer experience.

I am actively contributing to open source projects and leading the development of Catalyst, the open-source PyTorch framework for Deep Learning R&D.

In my spare time, I enjoy writing posts or give talks on ML-related topics.

Interests

  • Sequential Decision Making
  • Process organisation
  • Psychology

Education

  • MSc in Math and Computer Science

    Moscow Institute of Physics and Technology

  • BSc in Math and Computer Science

    Moscow Institute of Physics and Technology

Catalyst

Catalyst is a PyTorch framework for Deep Learning Research and Development.

It focuses on reproducibility, rapid experimentation, and codebase reuse so you can create something new rather than write yet another train loop.

Break the cycle - use the Catalyst!

Star

Get Started View Documentation

Latest release

Course

Week 1: Deep learning intro

Week 1: Deep learning intro

Deep learning – introduction, backpropagation algorithm. Optimization methods. Neural Network in numpy.

Week 2: Deep learning frameworks

Week 2: Deep learning frameworks

Regularization methods and deep learning frameworks. Pytorch basics & extras.

Week 3: Convolutional Neural Networks

Week 3: Convolutional Neural Networks

CNN. Model Zoo. Convolutional kernels. ResNet. Simple Noise Attack.

Week 4: Object Detection, Image Segmentation

Week 4: Object Detection, Image Segmentation

Object Detection. (One, Two)-Stage methods. Anchors. Image Segmentation. Up-scaling. FCN, U-net, FPN. DeepMask.

Week 5: Metric Learning

Week 5: Metric Learning

Metric Learning. Contrastive and Triplet Loss. Samplers. Cross Entropy Loss modifications. SphereFace, CosFace, ArcFace.

Week 6: Autoencoders

Week 6: Autoencoders

AutoEncoders. Denoise, Sparse, Variational. Generative Models. Autoregressive models.

Week 7: Generative Adversarial Models

Week 7: Generative Adversarial Models

Generative Adversarial Networks. VAE-GAN. AAE. Energy based model.

Week 8: Natural Language Processing

Week 8: Natural Language Processing

Embeddings. RNN. LSTM, GRU.

Week 9: Attention and transformer model

Week 9: Attention and transformer model

Attention Mechanism. Transformer Model.

Week 10: Transfer Learning in NLP

Week 10: Transfer Learning in NLP

Pretrained Transformers. BERT. GPT. Data Augmentation in Texts. Domain Adaptation.

Week 11: Recommender Systems

Week 11: Recommender Systems

Collaborative Filtering. FunkSVD. Neural Collaborative Filtering.

Week 12: Reinforcement Learning for RecSys

Week 12: Reinforcement Learning for RecSys

Reinforcement Learning. DQN Algorithm. DDPG Algorithm. Wolpertinger.

Week 13: Extras

Week 13: Extras

Research & Deploy. Config API. Reaction.

Recent Posts

Projects

Kittylyst

Kittylyst

A tiny Catalyst-like experiment runner framework on top of micrograd.

Deep Learning with Catalyst

Deep Learning with Catalyst

Deep Learning with Catalyst.

ML workflow

ML workflow

Checklist and questions to make ML right.

Catalyst codestyle

Catalyst codestyle

Accelerated Python code formatter

RL Intro

RL Intro

Brief introduction to Reinforcement Learning and Recommender Systems.

Alchemy

Alchemy

Experiments logging & visualization

NeurIPS RL 2019 challenge

NeurIPS RL 2019 challenge

2nd place solution for NeurIPS RL 2019 challenge.

Reaction

Reaction

Convenient deep learning models serving

Catalyst detection

Object detection with Catalyst

Catalyst segmentation

Image segmentation with Catalyst

Catalyst classification

Image classification with Catalyst

Catalyst.RL

Catalyst.RL

Catalyst.RL: A Distributed Framework for Reproducible RL Research

Catalyst

Catalyst

Accelerated deep learning research and development

NeurIPS RL 2018 challenge

NeurIPS RL 2018 challenge

3rd place solution for NeurIPS RL 2018 challenge.

OpenAI Retro Contest

OpenAI Retro Contest

4th place solution for OpenAI Retro Contest.

NIPS RL 2017 challenge

NIPS RL 2017 challenge

3rd place solution for NIPS RL 2017 challenge.

Contact