Hi, I’m Alexey Zubritskiy!
Software Engineer, Architect
I’m a passionate programmer. I mean it. I love challenging tasks and difficult problems. I started writing code even when I had no device to execute it and never stopped since then. I enjoy overcoming hardware and software limitations. I have fun in learning a new technology inside and out, adopting its way of doing things, and making exceptional new products with it.
I’ve replaced the heart of a complex system while it was running full throttle. I’ve designed new systems from scratch. I’ve dealt with various environments—from calculators and phones to clouds and supercomputers.
Programming is not only my job and my hobby. It is my lifestyle.
Profile
Full name
Alexey Sergeevich Zubritskiy
Born
10 Feb 1984 - Sverdlovsk, USSR
a@zubr.me
Phone
+7 904 544 84 91
Website
zubr.me
Skills
Machine learning & NLP
1 year experience, current priority
Tensorflow ML theory Stanford NLP tools
And also
- AMR
- RDF
- word2vec
- MATHLAB
Frontend
10+ years experience
JavaScript HTML CSS AngularJS 1.X BackboneJS
Side dishes
- JSON
- AJAX
- HTTP
- JQuery
- Underscore.js
- LESS
- SASS
- Jade
- JSDoc
- Markdown
- Grunt
- Mocha
- Jasmine
- Wintersmith
- Dust.js
- Cross-browser
- Async
Server side
9 years experience
Scala Java .NET Python Node.JS
Including, but not limited to
- Haskell
- Akka
- sbt
- ScalaTest
- Ant
- Maven
- JUnit
- ASP.NET
- C#
- VB.NET
- REST
- OOP
- Concurrent computing
- Dependency injection
Mobile
2 years experience
Objective-C & iOS Java & Android Titanium Appcelerator
Plus
- Cross-platform development
- AppStore publishing
Cloud
4 years experience
Amazon EC2 Amazon S3 OpenStack
As well as
- Amazon RDS
- Amazon RedShift
- Amazon DynamoDB
- Other AWS services
Data
10+ years experience
More keywords
- Database design
- Transactions
- Stored procedures
- ORM
- Serialization
- JSON
- Protobuf
- XPath
- DocBook
- SQL Server
- Oracle
- SQLite
- Blazegraph
- SPARQL
- NoSQL
Tools
my usual workplace
Git IntelliJ Idea Eclipse Linux
Less used tools
- Jira
- SublimeText
- Visual Studio
- Windows
- TFS
- bash
- curl
- MacOS
- xCode
Work Experience
Firmshift
Machine learning engineer, Architect - November 2015 - Present
Machine Learning powered initiative aimed at advancing current state of the art in deriving meaningful insights from unstructured data: Natural Language Processing, Abstract Meaning Representation, Machine Reasoning, and Semantic Unification.
- Developing architectures, software solutions, and Machine Learning models for various domains, including Biotech and Public Services.
- Developing apps for supercomputers and clusters based on Massively Parallel Processing model (Python, TensorFlow)
- Developing and integrating natural language processing algorithms (Scala, Akka, Protobuf, Python)
- Researching approaches for meaning extraction and merging (AMR, Blazegraph)
- Working on Data Visualisation solutions (HTML5, JavaScript)
Elsevier
September 2013 - July 2016
PharmaPendium - Foresee the Best Drug Development Strategy, model Drug Behavior In Vivo.
PharmaPendium offers dedicated data modules that provide insights and information on the critical focused areas of drug development, drug safety, ADME and drug-drug interactions. Challenges are inevitable on the path to regulatory approval of drug candidates. Successful drug development involves anticipating those challenges as early as possible. Comparative and up-to-date information is needed to verify pharmacokinetic and experimental data, anticipate species-specific issues in translating in vitro and in vivo studies, predict potential drug–drug interactions and design trials to best preclude regulatory issues.
To successfully perform critical functions such as pharmacovigilance, generating high-quality systematic reviews in support of evidence-based medicine and drug and medical device efficacy studies, pharmaceutical and life sciences organizations require a wide array of biomedical evidence. Embase increases the discovery of biomedical evidence to support critical life sciences functions, delivering relevant, up-to-date biomedical information to the global biomedical research community.
Frontend developer, Architect
- Creating architecture and step-by-step plan of migration from JSP-based frontend to Single-page application
- Developing Single-page application frontend, performing migration (AngularJS, BackboneJS, HTML, CSS, JS, Grunt, Ant)
- Developing frontend for new modules
- Integration frontend build toolchain into existing project build system (Node.JS, Grunt, Ant, Jenkins)
- Introducing and supporting newly joined developers
- Introducing and enforcing code style and documentation standards (JSHint, JSDoc)
ScienceDirect is Elsevier’s leading information solution for researchers, teachers, students, health care professionals and information professionals . It combines authoritative, full-text scientific, technical and health publications with smart, intuitive functionality so that you can stay informed in your field, and can work more effectively and efficiently.
Full stack developer
- Developing documents downloader applet (AngularJS, HTML5, CSS)
- Developing documents downloader service (Java, REST, Amazon Redshift, Amazon S3)
Paytogether (own project)
Mobile Application Developer - January 2013 - August 2013
Track shared expenses
Do you travel in group? Do you make a party? Do you have some other situations when someone is paying for everyone? And do you know who owes whom at the end?
“Pay together” is a useful notebook tracking group expenses and everyone’s balance. Your group can repay fully at any moment. Also you can repay for a single member only if he decides to leave the party.
Someone pays for everyone, everyone pay for someone, several members pay for some other members: these and all other combinations are supported.
Additionally you can control everyone’s and group expenses at any period to spend money wisely.
- Developing iOS application (Objective-C, SQLite)
- Developing Android application (Java, SQLite)
- Developing ORM library for android platform (Java, SQL)
- Publishing to AppStore
Hunite
November 2007 - December 2012
The Hunite Solution base offers employee and manager self service HR features. Through a web app, customers can purchase one or more solutions and offer these to their workers. The Hunite solutions read employee data from, and write changes back to, the companies ERP/HR systems. Supported systems included: SAP, Afas Profit, Raet Beaufort and Peoplesoft.
Each solution could be rolled out quickly as customers only had to set their parameters and options.
For all workers:
- Hunite Central: starter solution, offering employee integration and the web portal
- Digital Personnel File: Maintain digital employee files in Hunite. Including scan solution.
- Personal Data Change: Change your personal data such as names, address, relationships, etc.
- Front Office: Easily ask questions to HR, ensuring they and up at the proper HR specialist.
- Expense Claim: Upload your receipt or send it in by paper, fill out the form and get your money back.
For managers:
- Absence: Easily report and manage employee illness and maternity leave.
- Employment Change: Change your employees Employment data (promotions, salary, etc)
- In/Out Service: Easily register new or end existing employment records for your employees.
Full stack developer
- Developing HRM web application (VB.NET, ASP.NET Web Pages, jQuery, AJAX, JavaScript)
- Developing database (MS SQL, T-SQL SQL reports)
- Doing researches, technical and architecture design, proof-of-concept
- Interviewing, introducing and supporting new team members
Mobile developer
- Developing cross-platform HRM mobile application (Titanium Appcelerator, JavaScript)
Documentation manager
- Creating software, process and structure for product requirements system (DocBook, XSLT)
- Managing team of requirements creators
SoftDev Ltd
Full stack developer - February 2005 – November 2007
- Developing on-line Enterprise Project and Portfolio Management System (ASP.NET, C#, SQL Server)
Interests
Functional and reactive programming
Haskel, Scala, Akka, React.js
Composition of High order functions folds and ties data flow in unbelievable ways. It is just a new world where you learn programming from scratch.
Low-level software
Rust
In contrast, Rust is the way to safe and blazing fast calculations on almost bare hardware. I’d like to make a tiny ADs filtering proxy server capable to work on routers and even 3g modems.
Pet projects
Paytogether iOS app
Adblock2privoxy filters converter written in Haskel
SublimeText plugin on Python
And other projects on projects.zubr.me
Articles
Hough Transform on GPU (in Russian)
Fast Late-bound Calls with Generics in .NET
Straight way to create ASP.NET user controls library
And travels
travel.zubr.me
Education
Coursera
Machine learning
Principles of Reactive Programming
Functional programming in Scala
Ural State Technical University
Data processing and control systems
Thank you!
We are talking about how to deal with a problem that has no solution.”