Pavel Novichkov, Developer in Moscow, Russia
Pavel is available for hire
Hire Pavel

Pavel Novichkov

Verified Expert  in Engineering

Applied Mathematics Developer

Moscow, Russia

Toptal member since February 17, 2016

Expertise
Machine LearningPythonLinuxNumPyC++AlgorithmsNLPMySQLGit
Bio

Pavel is an excellent analytical problem solver with a passion for science and programming. He has a strong working knowledge of algorithms, statistics, and machine learning. Pavel's particular strength is his ability to apply mathematical tools and physical intuition to real-world problems and transform ideas into high-quality software.

Portfolio

WorldQuant
Bash, Python, C++
Yandex
MySQL, MapReduce, Python
Datadvance & Institute for Information Transmission Problems
MATLAB, Python

Experience

  • Python - 5 years
  • Applied Mathematics - 5 years
  • Machine Learning - 3 years
  • C++ - 1 year

Availability

Part-time

Preferred Environment

Subversion (SVN), Git, IPython Notebook, Jupyter Notebook, Emacs, Linux

The most amazing...

...thing I've coded is an algorithm for estimation of rare event probabilities that works 1,000 times faster than classical methods.

Work Experience

Quantitative Researcher

2015 - 2015
WorldQuant
  • Developed several algorithmic computer-driven trading strategies for equity markets.
  • Implemented automation tools in Python and Bash to simplify strategy backtesting and statistics aggregation.
  • Worked on a C++ library for multivariate time series analysis (vectorized mathematical operations, statistics, technical indicators) and uniform handling of various data types.
Technologies: Bash, Python, C++

Developer

2014 - 2015
Yandex
  • Optimized the ads allocation algorithm for Yandex search engine results page.
  • Worked on improvements of machine learning algorithms for click-through and conversion rate prediction.
  • Analyzed large datasets with Yandex MapReduce.
Technologies: MySQL, MapReduce, Python

Research Intern

2011 - 2012
Datadvance & Institute for Information Transmission Problems
  • Conducted research in the areas of uncertainty quantification and dimensionality reduction.
  • Designed and developed an algorithm for effective estimation of small failure probabilities (based on Gaussian regression and importance sampling Monte Carlo method).
  • Performed comparative analysis of existing methods for non-linear dimensionality reduction, and worked on development of novel algorithms.
Technologies: MATLAB, Python

Experience

Yandex Banner System

https://advertising.yandex.com/
I worked on increasing relevance, click-through rates, and conversion rates of ads placed on Yandex search results pages.

COMET Track Recognition

This Kaggle competition was held as a part of Summer School on Machine Learning in High Energy Physics. The task was to determine which points (gathered from detector measurements) belong to signal track. I applied a sophisticated mathematical trick (geometric inversion + Radon transform) to extract important features. It allowed me to achieve 0.9969 ROC AUC score with a simple gradient boosting classifier.

Kalah (Bantumi) Board Game

An application for playing African Kalah board game against the computer. The application was written in C++ and used Qt framework for the graphical user interface. The AI player was implemented with alpha-beta pruning algorithm and several evaluation functions.

Education

2013 - 2015

Master's (with Honors) Degree in Mathematics and Mathematical Physics

Higher School of Economics - Moscow

2008 - 2012

Bachelor's (with Honors) Degree in Applied Mathematics and Physics

Moscow Institute of Physics and Technology - Moscow

Skills

Libraries/APIs

Scikit-learn, Pandas, NumPy, Matplotlib, SciPy

Tools

Emacs, LaTeX, IPython Notebook, Subversion (SVN), Git, Mathematica, MATLAB

Languages

Python, C++, SQL, Bash, C

Paradigms

MapReduce, Object-oriented Programming (OOP)

Platforms

Linux, Jupyter Notebook

Storage

MySQL

Other

Applied Mathematics, Machine Learning, Scientific Computing, Data Science, Algorithms, Data Structures, Natural Language Processing (NLP), Convex Optimization, Generative Pre-trained Transformers (GPT)

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring