Vsevolod Bazhan Email & Phone Number

- Increased Apache Impala data warehouse stability by 70% (measured by number of alerts) and query performance by 40% (measured by 95th percentile query runtimes) by reevaluating node roles to free up existing memory, setting guardrails such as query timeouts, memory and disk usage limits, and implementing efficient data compaction techniques (addressing the “Small Files Problem").- Built a new analytical platform using an open source version of Apache Airflow. The platform components were deployed on an existing Kubernetes cluster and Airflow tasks were run in dedicated pods on AWS spot nodes. This approach reduced costs and enabled greater resource scalability.- Increased customer confidence in the data being delivered by building a custom SQL-based data quality and SLA notification system to detect and respond to data quality issues.- Fostered a culture of open communication by constantly reporting issues and features in a public Slack channel, leading to improved transparency and cross-team collaboration.- Developed and maintained near real-time streaming pipelines ingesting data from Kafka topics and ETL pipelines extracting data from external APIs and operational Postgres-based databases.- Actively maintained and improved a legacy in-house data processing solution written in over 100,000 lines of Python code.

- Helped students learn the data engineering profession: tested their projects and gave feedback. Shared experience, best practices, and an inside look with students.- Gained valuable experience in communicating with newcomers, developing the skill of giving eco-friendly feedback.

- Built and maintained ETL pipelines scheduled using Apache Airflow.- Performed complex scraping and parsing of Excel-based reports and external APIs.- Created datasets (tables, views and materialized views) for ClickHouse-based data warehouse.- Developed a parsing solution for Kafka-served streaming complex data packaged in gRPC format. The data was provided by a third party and the schema was generated automatically, which presented a challenge in making sense of the data and how it could be parsed.

Extended functionality of the chatbot builder platform through the development of REST APIs on Django and Flask.

Gained valuable experience working with a loosely documented proprietary game engine.