Pedro Zacharias

Part of the Data Domain Team implementing Data Mesh principles, evolving from building robust data products to developing AI-powered solutions, collaborating with Backend Engineers, Product Analysts, and Data Scientists to deliver domain-oriented, decentralized data ownership.

• Built domain-oriented data products for Motors platforms serving 2M+ MAU and processing 10M+ behavioural events/day.
• Reduced daily pipeline runtime from 8h to 1h and cut AWS Spectrum costs by 80% by refactoring datasets (column optimization, table decoupling, idempotent Airflow pipelines).
• Operationalised Data Science workflows: refactored Python/Spark/SQL code, set up CI/CD (GitLab, Docker), provisioned infrastructure with Terraform, and integrated MLflow for reproducible tracking.
• Developed and maintained low-latency APIs (AWS API Gateway, FastAPI) handling 30K+ monthly requests, with monitoring via New Relic dashboards and alerts.
• Developed an internal AI agent transforming natural-language business questions into actionable answers, using LLMs, vector search, and RAG-style pipelines served through company-wide APIs and MCP tools.

Part of the Data Domain Team implementing Data Mesh principles across the company — developing data pipelines, refining data models, and driving data engineering best practices to foster decentralized data ownership and domain-oriented design using AWS, Trino, and Airflow.

• Developed entirely new domain-oriented data models to support vital domain analysis: modelled data from business needs, built batch-processing pipelines, and refactored unstructured processes into optimized models by decoupling large tables, making pipelines idempotent, and optimizing column partitions, distributions, and compression — reducing AWS Spectrum usage by up to 80% by syncing data from S3 to Redshift.
• Assisted Analytics and Data Science teams in setting up Data Infrastructure resources, including S3, GitLab, and Airflow instances; promoted decentralized Data Ownership through Data Contracts, Data Quality alerts, Catalog Management, and governance practices.
• Acted as Scrum Master for a team of 9 Data Engineers — facilitating Scrum ceremonies, organizing JIRA to team standards, implementing an Incident & Response Workflow with PagerDuty, and introducing OKR Monitoring in Jira.

Part of the team responsible for building the European Data Ecosystem from scratch for over 400 employees using the most up-to-date cloud technologies in Azure.

• Delivered critical data pipelines and reports for the Underwriting domain, helping achieve the team's objective of delivering over 50 reports in 1 year.
• Implemented data pipelines in a Data Lake + Data Warehouse architecture and built Power BI reports to present the final data.

Professional internship at the leading consulting company for ERP PHC, working as a consultant.