SSP for US TOP-3 AdTech player
A major US based SSP (supply side platform) with complex custom core functionality, lightning fast performance and unique auction PMP algorithms with features enabling a possibility for a publisher to make direct deals with advertisers. The platform featured an advanced real time statistics.
Advanced targeting system
An advanced cookieless advertising solution based on page contents which employs web crawling, dynamic databases, proprietary machine learning algorithms and framework semantic scoring. It categorized over 1 million web sites, analyzed and placed them under specific categories for accurate targeting. We integrated several 3d party DMPs for additional info on web pages in various categorie.
Mobile advertisment programmatic
A platform for mobile user acquisition & re-targeting. It connects to dozens of different Supply Side Platforms, processes 1.5M requests per second, manages data of billions of users in real time, uses machine learning models to predict user behavior, and provides customized reports.
For a prominent US based Edtech company we have built an advanced pipeline solution which utilizes Postgresql, Apache Kafka, KSQL and Snowflake. The project involved high level architecture, design of the infrastructure, hands on deployment and configuration and then we provided level 3 support escalation.
FinTech mobile application for loans company
A mobile app for a finance company. It manages user data, tracks their loans and payments. Users can register and apply for a loan, review payment history and make payments for the current month or the whole loan. The app automatically evaluates the user's car cost when they apply by using a data model algorithm.
BI system for automobile market
A platform for the second hand vehicle market. It acquires data from the major marketplaces of second hand cars, aggregates and cleans information. Then it tracks the dynamics of price changes and uses data models to forecast prices for the cars.
In recent project we tackled technical debt and expenses. By analyzing the architecture, we avoided a full rewrite. MongoDB's limitations led us to Aerospike and AWS Athena. Migrating data improved performance and cut server needs. Server costs dropped by 30%. We scaled to 1 billion requests per day.