Andrey Zaytsev, Developer in Chicago, IL, United States
Andrey is available for hire
Hire Andrey

Andrey Zaytsev

Verified Expert  in Engineering

Software Developer

Location
Chicago, IL, United States
Toptal Member Since
September 21, 2018

Andrey is a full-stack, mobile, and systems developer who's worked at top companies such as Google, Dropbox, Apple, and Jump Trading. He's also completed his master's degree in computer science at the University of Illinois at Urbana-Champaign with a focus on machine learning. He possesses in-depth knowledge of C++ and Python, having worked with them for over seven years, and has experience developing in Java, JavaScript, and Swift.

LinuxC++PythonGitJavaAndroidiOSObjective-CMachine LearningMacOSMySQLPostgreSQLNVIDIA CUDATest-driven Development (TDD)Unit Testing

Portfolio

Google
Protobuf, Java, Python, C++
Jump Trading
NVIDIA CUDA, CMake, Splunk, NumPy, Python, C++17, C++
Barqo
Piwik, HTML, JavaScript, MySQL, PHP

Experience

Linux - 10 yearsC++ - 10 yearsPython - 10 yearsGit - 7 yearsJava - 7 yearsAndroid - 4 yearsiOS - 2 yearsJavaScript - 2 years

Availability

Part-time

Preferred Environment

SSH, Emacs, Visual Studio Code (VS Code), Atom, Git, Linux, MacOS

The most amazing...

...project I've worked on consisted in shipping a full-stack distributed system to perform admin actions at Dropbox.

Work Experience

Software Engineer

2018 - PRESENT
Google
  • Improved the Google Search release process by making quality evaluation more stable.
  • Defined metrics and implemented monitoring for an environment with over 1,500 service instances.
  • Developed integration tests and used them as a basis for automatic environment updates.
  • Created a universal reusable configuration for Google Search servers across non-production environments.
  • Simplified a Google Search server startup experience by taking advantage of a common API in two distinct environments.
Technologies: Protobuf, Java, Python, C++

Software Engineer

2016 - 2018
Jump Trading
  • Designed and implemented scalable low-latency systems for on-disk data storage.
  • Collaborated with trading teams to enhance a market simulation by redesigning the event pipelines.
  • Improved latency tracking to achieve an accurate simulation across more than 20 exchanges.
  • Worked with quantitative analysts to develop a distributed market simulation system using CUDA.
  • Developed a library to connect to strategies and trade on three well-known exchanges.
  • Implemented a custom query data analytics pipeline in Splunk to display real-time trading metrics.
Technologies: NVIDIA CUDA, CMake, Splunk, NumPy, Python, C++17, C++

Software Engineer

2015 - 2016
Barqo
  • Developed and deployed an intelligent boat search system that utilizes machine learning.
  • Redesigned the front-end to provide a more streamlined user experience.
  • Used web analytics tools to analyze user behavior in production and improve workflows and UI.
  • Refactored the entire back-end codebase to allow for better scalability, reusability, and readability.
  • Introduced regression tests as well as new workflows to ensure new changes do not break existing functionality.
  • Led new feature development at a startup with only two developers.
Technologies: Piwik, HTML, JavaScript, MySQL, PHP

Software Engineer

2015 - 2015
Dropbox
  • Shipped a distributed system for performing admin actions on the scale of 100,000 users.
  • Integrated the project with the internal systems and provided support to other teams.
  • Implemented remote procedure calls using Protobuf to create a robust and scalable system.
  • Designed and developed both the front end and the back end in line with the internal company infrastructure.
  • Collaborated with five different teams to test the solution and roll it out into production.
Technologies: Protobuf, React, ECMAScript (ES6), JavaScript, Django, Python

Software Engineer

2014 - 2014
Apple
  • Enhanced the end-user experience by improving the iOS and Mac OS X input methods.
  • Created a uniform design to provide continuity across mobile and desktop platforms.
  • Collaborated with multiple frameworks teams to test the features and get end-user feedback.
  • Ensured backward compatibility with a wide variety of legacy hardware.
  • Verified the features by utilizing integration, unit, and E2E tests.
Technologies: Cocoa Touch, Cocoa, Swift, Objective-C

Distributed System to Perform Admin Actions

http://dropbox.com
This system allows Dropbox admin team to respond to real-time threats such as spreading malware or sharing of various illegal content in a matter of seconds. The asynchronous architecture allows the admin to perform actions on 100,000 users with a single click of a button and all contained within a user-friendly UI.

Low-latency On-disk Data Storage Solution

I built this scalable system in C++17 to allow for maximum performance when working with various market data. The architecture of the on-disk DBMS allows to bypass the overhead of traditional systems such as MySQL and is specifically optimized based on the needs of trading and latency teams.

Faster Apprenticeship Learning through Inverse Optimal Control

I developed a novel scalable approach to performing apprenticeship learning using state-of-the-art deep learning research. The new method uses a matrix formulation to optimize the entire system as a whole, in contrast, with the traditional iterative approach. After implementing the technique using Python and TensorFlow, I was able to see a significant increase in the performance and scalability of complex machine learning problems. This research effort resulted in my master's thesis.

Balancing Trade-offs in Test-suite Reduction

http://mir.cs.illinois.edu/marinov/publications/ShiETAL14ReductionEvolution.pdf
A publication in the 22nd ACM SIGSOFT International Symposium on the Foundations of Software Engineering.

Recourser Project

http://github.com/andzaytsev/recourser
A full-stack web application designed to allow university students plan their college careers. The back-end uses a real University of Illinois course database that's updated in real time. The features include an intuitive drag-n-drop UI as well as a total credit hour calculation. The project uses Python Flask for the back-end and jQuery with a jQuery UI for the front-end.

Languages

C++, Python, Java, Objective-C, JavaScript, CoffeeScript, R, Scala, Swift, ECMAScript (ES6), C++17, PHP, HTML, Octave

Tools

Git, Splunk, Mercurial, Subversion (SVN), LaTeX, Atom, Emacs, CMake, Piwik

Platforms

Linux, Android, MacOS, iOS, NVIDIA CUDA, Visual Studio Code (VS Code)

Other

Machine Learning, Software Architecture, SSH

Frameworks

TestNG, Django REST Framework, Flask, Django, Android SDK, Cocoa Touch, Cocoa, Redux

Libraries/APIs

Protobuf, React Redux, Theano, TensorFlow, NumPy, React, jQuery UI, jQuery

Paradigms

Test-driven Development (TDD), Unit Testing, Functional Programming, Agile

Storage

MySQL, PostgreSQL

2017 - 2017

Master of Science Degree in Computer Science (Machine Learning)

University of Illinois at Urbana-Champaign - Urbana-Champaign, IL, USA

2013 - 2016

Bachelor of Science Degree in Computer Science

University of Illinois at Urbana-Champaign - Urbana-Champaign, IL, USA