At Roku, I lead data-platform architecture for advertising datasets processing 200M impressions and 50M conversion events daily, while maintaining 100+ production Iceberg tables. I've also redesigned GDPR download and deletion workflows, increasing intake capacity from 20 to 1,000+ requests daily and reducing deletion fulfillment from 90+ days to 18 days.
Previously at Groupon, I built personalization services delivering 40M emails and 60M push notifications daily at 5 ms p95 latency, migrated 2,500 Spark jobs to Dataproc, and cut a datastore's annual cost from about $2M to $200K. I work across Python, Scala, Java, Spark, Iceberg, cloud data systems, and production infrastructure, with a focus on reliable systems that teams can operate safely.
