At MoniepointMFB, I introduced dbt as the team's transformation layer, adding tests, lineage, and documentation; pipeline uptime is 97%. I also built a real-time reverse ETL pipeline in Apache Flink that denormalizes application data into Elasticsearch for the customer support platform.
At Alerzo, I was one of three engineers moving the data platform to self-managed ClickHouse, Docker, and Python pipelines, and I led the GCP side of the migration. Monthly costs fell by about 85%, to roughly $200–300; I also designed a ClickHouse load tool and maintained ETL workflows. Earlier, at Anadata, I led analysis that traced address verification errors to case sensitivity, improving accuracy by 10%.
