Big Data Engineer
Jul 2021 – Present- Tuned Spark jobs over 3 PB of network telemetry with salting and adaptive query execution, cutting EMR costs 48%
- Built a Flink streaming pipeline processing 2M events per second for real-time outage detection
- Moved Hive tables to Apache Iceberg, enabling time travel and cutting full-table rewrites
- Partnered with network operations to deliver dashboards that shortened outage detection from 20 minutes to 2