At Oracle, I own 14 production pipelines generating an 80TB+ multimodal training corpus, and I’m the sole maintainer of the Kafka layer used by a 12-person team.
I reduced validation failures from 12% to 4% of daily volume by adding typed validation gates and error-class routing. I also built internal REST and tooling APIs used by five teams.
At Reva Group (Rentola), I rebuilt the listing-collection platform across 1,400+ feeds and sources, cutting a full refresh from 16 hours to under three minutes.
I served the cleaned data through REST APIs backed by PostgreSQL and Redis, and tuned an 800M-row PostgreSQL table to reduce p95 query time from 3.8 seconds to 160 milliseconds. I’ve also led a four-engineer pod and supported peers through pairing, review, and mentoring.
