Projects: OLAP backend, ETL, ML recommendation systems, R&D, other internal projects

Achievements:

— created ML product recommendation pipeline that was crunching terrabytes of raw data and served up to 1k product recommendation requests per second with response time SLA < 20ms

— created Recommendation Engine, that managed 100+ user-defined hierarchical models and served up to 1k rps with model response time under 20ms

— partially designed and worked on Analytics pipeline, that provided ETL and OLAP capabilities for billions of events per day, with petabytes of historic data and real-time reporting capabilities

Technologies: java, scala, kotlin, python, spark, spark ML, vertx, kubernetes, docker, AWS, mysql, vertica, clickhouse, cassandra, kafka, etc.

— redesigned and implemented OLAP backend, with 10x speed for some queries and extendable architecture

— created a tech stack from scratch and then successfully lead a team of 3+ people building and delivering Playable ads (a web-based replicas of mobile games)

— participated in development and optimization of an ML-based RTB bidder, a high-loaded service with billions of requests per day, reduced RAM requirements 4x

— as a part of a medium-sized team (6-10 software engineers) redesigned ETL pipeline for Ad Server to handle the increased traffic (up to one billion impressions per day)

— designed snowflake-like scheme in Postgres, sql-generating backend service, real-time analytics

— using extensive knowledge of the whole system was able to fix the most tricky bugs, optimize performance of critical parts, and provide L3 support

— Designed and developed a corporate CMS based on Zend Framework

— Implemented web-based hotel reservation system for partner company

— Created big number of smaller websites for partner companies

— Successfully released several versions of the project

— Implemented trigger-based change history logger

— Improved development process in several ways