Project is a Customer Relationship Management system. It is used for automation lottery ticket orders. The main objective of the project is to communicate services intended for these goals through their API’s. Currently the project is used in European countries such as Germany, Denmark, Finland, Sweden. Each of them has their own CRM (copy) with its own features and settings.

The main goal of the project is to find connections between di�erent entities in social networks . They use a lot of connectors to perform search in such sources as Facebook, Instagram, Linkedin, Twitter etc.

Eckard Enterprises is a conglomeration of private, family owned firms in the energy space and focused on the U.S. oil and gas industry. Eckard has a wealth of knowledge and expertise they o�er to private investors who seek to develop a successful energy portfolio. The Eckard Team is tirelessly committed to identifying, acquiring, developing, and maturing various energy-related assets for our partners. EE believes that owning tangible energy assets is one of the safest, long-term investment strategies available in today's investment climate.

CNXION is easily configurable online software that creates ‘matching’ solutions for any type of application domain in which there are two parties, an ‘o�wner’ and a ‘demander’.

RooM the Agency is a new mobile-driven community that brings together street commercial mobile visionaries with a traditional stock photography platform. On the back part, it’s intended to give the ability to manage the platform, keep statistics, create sales reports, and control payments.

Proven is a private platform where your portfolio companies can contact a select group of professional services providers. You know who you want to recommend, and Proven helps you to connect the highest quality professional services suppliers with your portfolio companies. Finding the right service providers is incredibly important for new companies.

The main task of this project is automatization of procurements in Google Express. All details with new orders stored are uploaded from xls file. After that, the crawler starts and passes through all the steps to create new orders. Selenium WebDriver was used there. The tool is intended mainly for creation test automatization scenarios. But in this case, use was more than justified because Google has good scraping protection and Selenium has the ability to walk through the pages almost like a regular user

It’s intended to get data from several Chinese sites with real estate and aggregate all gathered data in one place. For this project was deployed Scrapyd (the system of control and launch for scrapers)

A tool for the analytics department of an Israeli startup. The tool was designed to compile analytics for ads launched on sites. It used such technologies as Selenium, Celery, Django, Django Rest, Redis, PostgreSQL. We analyzed both text and video advertising. Here we had to deal with very large amounts of data. More than 20 servers are involved in the project. Several million pages a day were visited, each of which generated megabytes of analytical information. The bottom line was that the workers had to emulate the behavior of browsers and record how the ad is displayed, what kind of advertising, whether there is autoplay (sound / video). There were also written tools that tried to identify the entire chain of suppliers of this advertising, and build statistics. Ansible was used to control everything.

As part of data-mining, I participated in several large projects. The most large-scale was aimed at collecting information from a large number of online stores. On the project, Scrapy was used as the main scraping tool written on the Twisted asynchronous framework. Thanks to asynchrony, we managed to significantly speed up the execution of tasks. Some of the tasks on the project were solved using Selenium with PhantomJS, so I got experience of using these tools. The common queue for the distributed scraping system was first based on Redis. Then they switched to Amazon SQS. I had to rewrite the Scheduler and part of the middleware.

Project is designed to collect and analyze information from social networks such as Twitter, Instagram, Pinterest. I conducted this project independently and was engaged in deployment it on servers. It involved 5 servers. All of them perform the tasks of collecting and analyzing data to identify trends (directions that are also interested for users who make reposts from the analyzed tape). One of the servers is used to store all the results in the MySQL database. The project is built on an asynchronous Twisted framework that allows you to handle a large number of simultaneous requests. As a result, a large amount of intermediate data accumulated. In order not to clutter up the results base, and to speed up processing, a non-relational MongoDB database was selected for their storage.

When I worked with PHP I mainly worked with custom CMS - Universe. It was created by a customer. We worked on orders such as business card sites, e-commerce sites. Also we expanded and modified Universe CMS code.