Kanstantsin K.

Re-architected and stabilized the main Apache Flink pipeline (GDC) in the Unified Data Ecosystem (UDE STR) to handle multi-GBps vehicle telemetry traffic. Replaced an unstable Flink SQL implementation with a production-grade Flink Streaming API, rewriting 10k+ LOC. Delivered end-to-end large message ingestion (up to 64 MB), cold storage integration, and package merging, enabling seamless data migration across pipelines. Drove core architectural decisions, decoupled Flink configurations from the Kubernetes Flink Operator, added DFS support for Azure ADLS Gen2, and defined DevOps procedures for upgrading Flink in production. The pipeline has been rolled out globally (EU, US, CH), replacing legacy systems. Major Java backend and REST API changes, including validation layers and integration endpoints, to support new streaming and migration workflows.

As a core team member, contributed to the «TradeFinder» financial portal by migrating large data pipelines, scripts (python, matlab, sh), and web interfaces to newer versions, addressing security issues and supporting ongoing maintenance and improvements.

As a core team member, responsible for the design, implementation, and maintenance of microservices within HSBC's Payments Orchestration Layer framework, a key part of their high-value, high-volume payments system with 350+ production environments. The architecture used gRPC, Kafka, Kafka Streams, and Oracle in a resilient, low-latency mesh optimized for high throughput in a secure, regulated environment. Also provided Level 3 (L3) support.