Yaroslav Kopotilov, Data Scientist

Remote positions only

Toptal LinkedIn GitHub



Machine Learning

sklearn statsmodels lightgbm skopt tpot numpyro pytorch tensorflow

Data Engineering

pandas dask requests imaplib STOMP SQL SQLAlchemy AWS-S3 AWS-Athena Spark

Reproducible ML

mlflow environments Docker Git/GitHub Confluence


prefect Docker Ansible pytest pre-commit Git/GitHub ELK Sentry REST

Numerical Computation

numpy scipy numba jax

Data Visualization

jupyter matplotlib plotly dash flask waitress HTML

Industry Knowledge

Stack: Python | C++ | Bash

Other: Project management | Team leading | Mentorship

Work Experience

07/2020 - present: Data Science, Toptal, Remote
Building abstract models that add real value
01/2019 - 07/2020: Data Science, Vitol, London
Conquering energy markets with the power of ML
01/2019 - 06/2020: Model Validation, JPMorgan, London
Challenging pricing models for commodity options
05/2016 - 10/2016: Algorithmic Trading, Credit Suisse, London
Seeking for alpha in commodity and equity markets
02/2015 - 07/2015: Research, Novosibirsk State University
Writing a paper on image classification for one particular type of shapes

Academia Degrees

MSc in Financial Mathematics, Pierre and Marie Curie University
DEA El Karoui: top Financial Mathematics program in France
MSc in Applied Mathematics, Ecole Polytechnique
Cycle d'Ingénieur polytechnicien: top Engineering program in France
MSc in Mathematics and Computer Science, Novosibirsk State University
BSc in Probability and Statistics, Novosibirsk State University

Personal Projects

Algorithmic trading
A mid-to-high frequency ML-powered strategy (proprietary)
Data wrangling tools
Data pipelines with caching (link) & an efficient XML parser (link)
Interactive website
This very website powered by Flask + Dash (link)
Forecast Sales Kaggle Competition
A surprisingly simple model that won the competition (link)
Embeddings in Machine Learning
An article discussing what embeddings are and how to use them (link)
Popular classification algorithms
A comparison study using data from Titanic Kaggle competition (link)