At Arpeely, I own petabyte-scale data infrastructure on GCP, processing billions of daily events for product, ML, and business teams. I redesigned customer pipelines into a config-driven architecture, cutting costs by 30% and reducing table freshness from one hour to 20 minutes.
I also built an AI-powered PR analysis tool and an on-call briefing agent, and created pipeline monitoring that reduced data-starvation incidents from weekly to less than once a month. At oolo (acquired by AppsFlyer), I owned the data infrastructure and migrated the DWH to Redshift, while at eToro I replaced manual Excel workflows with ETL pipelines.

