Francesco Ardita

I led the end-to-end migration of a mission-critical ERP estate to Google BigQuery for a national strategic-infrastructure client. The platform runs on a 3-layer lakehouse pattern — raw, staging, curated — with Apache Airflow orchestrating ingestion and Type-1 SCD merge logic keeping the data auditable. The analytical SQL reports I authored replaced legacy ERP logic and were adopted for formal UAT sign-off by executive stakeholders.

I designed a generative BI platform that converts natural-language requests into live analytics dashboards. The architecture runs a multi-agent system on an open-source AI stack — each agent handles a specific task: schema mapping, query generation, chart selection. The concept was shortlisted for Leonardo's internal Innovation Award 2026.

I built advanced DAX measures for commercial KPI tracking that replaced a manual Excel workflow. Dynamic date-filtering logic and forecast/margin measures now integrate directly with the enterprise performance-management system, giving executive stakeholders real-time visibility into revenue initiatives.

I deployed a production-grade open-source BI stack for a public-sector healthcare client, replacing a proprietary tool that had stalled at pilot stage. I resolved the infrastructure and security issues that were blocking go-live and implemented consistent, brand-aligned dashboard theming across the entire platform.

I delivered a visual-analytics dashboard for regional healthcare-claims management, migrating the data to a cloud platform to enable drill-down risk reporting. The dashboard is now the primary tool for identifying claims anomalies and assessing financial exposure.

I contributed technical content to a defense-sector proposal for an Air Command and Control data platform. The work covered hybrid cloud/on-premises architecture, Zero Trust security patterns, and classified-data handling requirements aligned with NATO standards.

I built ETL pipelines that extract equipment and sensor time-series data from the manufacturing floor into a curated analytics layer on Databricks. The pipeline feeds anomaly-detection models that improved production monitoring and reduced unplanned downtime.

As Work Package Lead, I owned the cloud data flow end to end — from raw machine data ingestion to the ML models that detect non-conformance patterns. I coordinated cross-team delivery across semiconductor fabs in Asia, ensuring the models reached production on schedule.

I consolidated multi-platform client data into centralized databases and built custom reporting that identified billing discrepancies across systems. The reporting helped the client avoid €60,000 in revenue leakage — the kind of number that makes a pipeline pay for itself.

I built a cloud ETL pipeline that ingests social-media data into a data warehouse, feeding BI dashboards for digital marketing and e-commerce performance tracking. The dashboards unified campaign metrics that were previously scattered across four different platforms.