Leading Storage team developing and improving in-house distributed Exabyte-scale object, key-value and blob storage systems widely used at VK.

Managing master's degree program. Teaching courses on NoSQL and Highload systems.

Designed and implemented distributed partitioned replicated reactive ok.ru/music caching streamer. Improved hardware utilization, increased reliability and scalability and substantially improved 99.9% latencies.

Heavily contributed to the first release of becoming extremely popular open-source async-profiler.

Contributed to and maintained popular one-nio open-source project.

Designed, implemented and deployed core features and optimizations for internal Cassandra-based NewSQL storage. Substantially improved resource consumption and networking latencies.

Implemented media transport components (i.e. screen sharing data transport) of massive Sferum educational platform used by many schools in Russia.

Unified company wide onboarding process across all the backend, frontend and mobile teams.

Designed and implemented in-house Exabyte-scale AWS S3 compatible object storage becoming extremely popular at VK with buckets up to 10s of PB and 1M RPS production load.

Designed and implemented content quality control subsystem.

Unified technology stack and improved multi-location reliability of Auto.ru.

Participated in hiring activities, hardware resource planning, SLO enforcement.

Designed, reviewed and implemented critical internal components (libraries and microservices) used in offer aggregators (Auto.ru, Yandex.Realty, Yandex.Rabota, Yandex.Travel): offer indexing pipeline, search subsystems, billing, counters/throttlers, activity tracking, antifraud, content quality control, archival data processing, subscriptions, metering, monitoring, etc.

Was responsible for scalability and fault tolerance aspects at architecture/design reviews: Yandex.Rabota, Yandex.Travel, Reviews, UGC, etc.

Advocated and integrated Cassandra storage and Akka framework for building reactive concurrent services (presented in Russian at JPoint 2014). Introduced Hadoop ecosystem (HBase and Hive) for archival data processing.

Broadened code review and continuous integration practices in company development process.

Designed and implemented open-source embedded DBMS YoctoDB to replace Apache Lucene in Auto.ru/Yandex.Auto as a search index implementation -- service latency improved substantially (presented in English at CEE-SEC(R) 2016).

Participated in hiring activities, hardware resource planning, improvement of deploy processes, SLO enforcement.

Led migration from Subversion to Mercurial. Introduced Scala, Akka, Metrics, dependency management, continuous integration, code review practices.

Justified necessity for development of eternal archive of offer content and user activity data to make analytics possible.

Designed, implemented and reviewed internal services and components: user settings, subscriptions and notifications, picture processing and caching, Web crawling, archival data collecting and processing.

Transformed existing single-node indexing infrastructure into scalable distributed highly-available streaming indexing (the approach is described in the patent).

Advocated and introduced NoSQL storages: Cassandra, HBase, Redis, MongoDB -- substantially improved scalability and availability.

The results were presented in Russian at Yandex.Subbotnik and ADD-2012.

Led local functional community Saint-Petersburg Haskell User Group (SPbHUG).

Conducted thorough research on static analysis methods and algorithms.

Conducted thorough research on static analysis methods and algorithms.

The results are published.

Improved development processes: introduced Trac, Mercurial, unit testing.