Head Of Data Integration
CurrentMy current challenge is Provisioning of Trusted Data for AI Agent's. I approach this goal by leveraging Bank Semantic Layer: unified metadata layer, that consist of:• enterprise-wide Business Glossary• integrated Logical Model, interconnecting Business Glossary terms• Logical Model mapping to Physical Data Models of Data ProductsUsing this metadata our Lang Chain Applications built on top of Sber's proprietary LLM (GigaChat) are able to answer natural language requests:• find accurate data in enterprise Data Mesh and extract that data correctly• join and union datasets, extracted from different domains in accordance with request logicMy key achievement so far is creation of platform and framework for data providers to supply data consumers with integrated, reliable and accurate data. In order to achieve this goal I focus on following initiatives:• Unique metadata-based Data Platform with built-in Data Architecture framework and automated PDM, ETL and DQ scripts generation• Bank Data Mesh / Semantic Layer providing integrated data of highest quality• Data Architecture framework and tools, providing solution to Data Consistency, MDM, Data Accuracy challenges along with Data cost reductionOne more mission is to guide 200+ domain data engineers throughout the Bank in creation of their Data Products. To support this goal I provide:• 10+ weekly knowledge sharing workshops dedicated to all aspects of Data Product creation• Constant elaboration of learning materials and guides, training •. Data Integration Community.I lead a team of 60+ Data professionals, who elaborate architecture, develop Data platform and frameworks.Also I led a streaming Sber Ecosystem Data platform to enable NRT data provisioning for Recommendation Systems, NBO / NBA, Smart Targeting, Clickstream acquisition.All platforms and frameworks are in-house development based on open source: Hadoop, Spark, Java, HBase, K8s, Kafka, Spark Streaming, GreenPlum, Trino, ML Flow.