How it fits together
Event sources
8–16B events a day
S3 data lake
280+ TB
ETL pipelines
100+ GB a day · 690+ tables
Redshift warehouse
47 TB · tuned queries
Dashboards
TypeScript frontend · <35s queries
The challenge
The platform ingested 100+ GB and 8–16 billion events every day across 690+ tables, with 280+ TB in S3 and 47 TB in Redshift. Dashboard queries took up to two minutes, and a growing JavaScript frontend was getting harder to change safely.
What we did
- Worked as part of the offshore engineering team owning backend, data pipelines, and DevOps.
- Profiled the slowest Redshift queries and reworked table design, sort/dist keys, and query shape.
- Hardened the ETL that moves daily event volume from S3 into the warehouse.
- Led an incremental TypeScript migration of the frontend, page by page, without freezing feature work.