Tamara – Python, SQL, PySpark

An agency helping enterprises (like the Italian government, which tracked species in the Mediterranean sea with IOT) and startups (like fintechs and AI platforms) with their data.

• Refactored Python ingestion pipelines into production PySpark on Databricks, achieving a 1,000x performance improvement.
• Designed ingestion, transformation, and indexing pipelines delivering trusted datasets for LLM/RAG ML workflows.
• Built infrastructure-as-code for Databricks platform deployments using Bundles and Pulumi.
• Built scalable batch and real-time processing pipelines for large-scale geospatial datasets.
• Ensured data integrity and quality across multiple sources in collaboration with DS/ML teams.

British Petroleum is an oil and gas company with several focuses. Tamara worked on projects in Production & Oil, as well as Intelligence, Crisis, and Security.

• Architected data pipeline architecture end-to-end using Databricks Unity Catalog, Delta Live Tables, and Medallion architecture.
• Owned data model design across multiple greenfield client data platform projects.
• Optimized data workflows for performance, reliability, and cost efficiency.
• Established CI/CD deployment standards and monitoring for Databricks and Azure Data Factory pipelines.
• Integrated diverse, structured, and unstructured sources into governed Data Lakes.
• Enforced data integrity, quality, and security best practices.
• Led and mentored teams of up to 4 engineers.
• Collaborated cross-functionally with software engineers, data analysts, and product owners to deliver end-to-end data solutions.

A fintech startup in real-time payments. Tamara was responsible for data migrations, data quality, reporting for the clients, and stakeholders.

• Architected and maintained an analytical Data Warehouse using SCD Type 2 logic.
• Reduced BigQuery costs by 30% and improved data availability from 7 hours to under 2 hours.
• Led a zero-downtime migration of 2TB of data from MySQL to PostgreSQL using CDC-Debezium, Pub/Sub, Python, and Kubernetes.
• Refactored and migrated DBT models into BigQuery, ensuring logic parity and performance.
• Ensured data integrity, quality, and validation across operation vs analytics layers.
• Owned reconciliation monitoring and stakeholder reporting reliability in a live fintech data product environment.

A project for Intesa SanPaolo bank. Tamara was responsible for the Group Data Warehouse, which unifies payments, loans, real estate, legal, and natural-person transactions from Croatia, Slovenia, and Bosnia and Herzegovina.

• Designed, implemented, and optimized ETL pipelines and DWH architecture and data models using Oracle Data Integrator and Oracle DB for financial reporting and regulatory compliance.
• Collaborated with the business on data definitions and quality standards.
• Unified disparate data sources across Group entities into a single DWH.
• Automated advanced SQL queries for daily financial reporting (10x speed improvement).
• Optimized reconciliation logic execution time by 5x.
• Mentored junior engineers.