Data Platform: Changing Planes Mid-Flight

Sergey Mikhalev leads T-Bank's centralized data platform. About 230 developers build it for 15,000 monthly users. Shared rules, storage, and compute turn data into everyday infrastructure for product teams.

Sergey grew up in Gatchina, attended a border-service cadet school, and chose programming for its visible results. He applied for a C++ internship, but the opening had closed, so he started on .NET. Over eight years at the company he optimized an enterprise app, later worked on Java, and became a team lead. Larger datasets led him to Odnoklassniki, where his team owned the path from logs to experiments.

Sergey later led data platforms at Odnoklassniki, VKontakte, and VK's central analytics office. He left when technical decisions began arriving from above without discussion. T-Bank offered deep data adoption and a shared platform, together with a Greenplum ceiling where storage and compute could only be scaled together.

The new platform separates S3 storage from Spark and Trino compute, with Iceberg as its table layer. Tens of thousands of data feeds and scheduled jobs must move, domains must own resource costs, and hidden database reads must become explicit contracts. Sergey connects scale to federated ownership and personal growth to side projects and genuine interest.