Software engineer at e-com Yandex Search runtime

Leadership • Hired and managed the work of two people (an intern and a middle developer)Real Time Bidding ads system of ozon.ru • Designed and implemented ads attribution. Multiple micro-services to link buyer's events(clicks, views, to carts, etc) with seller's ads. Sellers use it to assess the effectiveness of ads campaigns. CPO attribution makes around half of total Ozon's ad revenue. [GoLang, Clickhouse, Redis] • Did 2nd generation of budget overrun control mechanism for sellers' ads campaigns. It reduces the number of incidents related to over-the-top bills for our clients in cases of tech issues [GoLang, Kafka, PostgreSQL] • Participated in creation of ads formats: media ads(banners, brand-shelves, video-banners), search engine advertising, external ads(google-shopping) • Running and analyzing many ab-tests to improve performance of ads campaignsOther • Talks at internal conferences (what ads attribution does, why use DI container in GoLang) • Crowdsource (common gitlab CI pipeline for GoLang libraries: tests, linters, secrets search, jira integration, git tags; shared internal libraries)

Pricing team• Implemented “PartnerAPI” to fulfill orders from other taxi services in case of lack of drivers supply [PHP7]• Created microservice for fiscal receipts generation at the end of trips - “Anfisa Checkovna” [GoLang]• Implemented Gitlab pipelines to run unit-tests, code coverage, linters, deploy to Kubernetes clustersSurge Pricing team• Implemented microservice for assignment of different surge models in cities of presence [Python3]• Re-designed and moved the legacy surge algorithm from AirFlow into Kubernetes CRON [Python3]Tech evangelism - kubernetes, unit-tests, linters, gitlab pipelines• Led the creation of the first Kubernetes cluster• Arranged a meetup of devops and engineer teams to discuss “what is kubernetes and why we need it?”

ForEx platform – Autobahn/RAPID, low latency exchange for High-Frequency Trading clients• Implemented price prediction quant project (research initiative). I checked one of quant’shypothesis about currency’s price prediction on historical data sets.Python (Jupyter, scipy), Matlab• Improving client pricing module, which is project’s core product. I participated in globalrefactoring, that significantly simplifies the process of adding new clients into the system.C++ 14, boost• Working on code performance optimizations and latency issues investigations:◦ One of the results is the tool, which automatically calculates the round-trip delay of the whole system based on our log records and Corvil’s traffic dumps. Python, Corvil◦ Also, found the reason of latency instability in production. I proposed the solution and proved it in UAT environment. C++ 14, rtdsc (ASM)