Owned delivery process for two projects (connectedTV, native players: ios, android) * create internal tools, tune infrastructure, improve video delivery metrics. * system to defect breaches, system coverage (front, back, mobile) to support content delivery, smooth playing, ads playing and reporting for mobile and TV platforms. * performance analysis/tuning on components. * tools for deeplinking, content delivery, ads payload, closed captions, configuration verification.

• Migrated test infrastructure to the cloud, reduced time to market metric by introducing multithreaded, scalable environment based on AWS infrastructure, Docker containers and Kubernetes orchestrator. Flexible power on demand, time savings on E-2- E routine increased by 35%. • Reduced tests flakiness from 25% to 7%, cleared up redundant routines to save execution time. Built client telemetry and logging system based on Splunk, crated dashboards for tracking quality metrics, content issues. Content issues spotted and fixed by producers, no broken piece of content live, performance metrics available for analyses from diff. regions and brands. • Designed and developed REST service to act as entry point to create low level test abstractions as user, movie, playlist, etc. Lift test layer up, on the level of REST service. No duplicate code, operates high level abstractions (scalability-ready by moving any backend routine to REST endpoints). • Unified scattered client code across several streams to one code base and one set of rules. Introduced DSL to use same abstraction in different sub projects and make re-use of existing functionality which helped to track integration problems, reduced time to assemble new test, reduced chances of getting new bug into test code. Onboarding of new brands requires only time to adjust UI layer. • Introduced infrastructure as a service conception: All required prerequisites to launch testing routine and prepare environments are integrated into CI/CD process using open source frameworks as Jenkins, Pipeline, Kubernetes, Docker. Tools: Python, Flask, Alchemy, Java, Spring, Retrofit2, Hibernate, Selenium Grid, Jenkins, Docker, Kubernetes, AWS, Splunk, TestNG, jMock, Maven.

* Functional manager for the group of 20 people. Built production unit ground up, organized people scattered across different teams into one unit being responsible for delivering services in media account. 3 projects delivered. * Started strategic people project to develop connection with universities, organized internal lab and courses to get best students trained and ready for production tasks, organized pool of managers to act as peers for best students. Benefit: More than 100 students hired, process was made company wide standard. * Designed and implemented course that teaches fundamentals of software engineering to make a conversion for people who is willing to make a career move. Benefit: 35 Engineers converted to TA after performance review procedure in a while. * Performed functional leadership for 20 direct reports. Benefit: Promoted corporate culture, defined career path and personal dev. plans.

* Built the tools to simulate end-to-end scenarios on various windows and linux platforms, utilize different setups of containers which talk to each other in defined network topologies. Simulate weak spots in highly distributed containerized environment. * Track system metrics, performance characteristics of custom linux images cooped with discovery services. * Tools to find data inconsistency issues between several sources: Tibco EMS, Tibco Rendezvous, Informix database and MSSQL server instances, rules of sharding, replication and scaling. * The platform could be described as micro services architecture decomposed by business capability. Tooling heavily relied on test doubles and consumer-driven methodology. Modular testing approach with ability to “hot swap” micro service with mimic wrapped into container. Tools and platforms: Parallels Virtuozzo Containers, CentOS/Ubuntu/Windows, JMS, Tibco EMS/ Rendezvous, MSSQL server, Informix, Java, Python.

chaos framework for distributed transaction rules, isolation levels, weak spots in 1 leader many followers architecture.