Byamba E.

Databricks data engineer with five years in production. I build the pipelines and lakehouse layers that a business can actually trust — and the AI applications that read from them. WHAT I DO • Design and run Databricks pipelines: PySpark, SQL, Delta Lake, Delta Live Tables, Workflows, Unity Catalog • Medallion (bronze / silver / gold) lakehouse design — what is kept exactly as it arrived vs. what represents current state • Incremental, idempotent loads: watermarks and MERGE on stable keys instead of full reloads, at tens of millions of rows • Data quality that fails loudly: count and key reconciliation between layers, so a partial feed stops the job instead of publishing a wrong number • RAG and LLM applications on top of the data — retrieval, evaluation, and the pipelines that keep the index current • CI/CD for data: Azure DevOps, Git, tests, monitoring WHERE I DO IT Right now I work in insurance, on claims and coverage data — regulated, person-level, and audited. I build the ingestion and modeling pipelines and the RAG applications for coverage advice and claims analysis that run on them. Before that: healthcare analytics and ETL on Google Cloud (see my Upwork history). PROOF • Databricks Certified Data Engineer — Professional and Associate • Rebuilt a production email-ingestion pipeline (Microsoft Graph API → bronze/silver/gold → AI consumer) from full reload to incremental: hours to minutes, with between-layer checks that catch missing data before release • Two products I built and run alone, end to end — data model, dashboards, auth, payments: TestLogicHub and DataLogicHub (both in my Portfolio below) • Every rated Upwork contract: 5.0 HOW I WORK I would rather tell you what I don't know than find out on your budget. I say what I will do, in plain English, and I explain design decisions so that non-technical people can agree on what a number means. Short daily updates, one question at a time, no surprises. GOOD FIT FOR Databricks pipeline builds and migrations · lakehouse design and audits · data-quality and reconciliation frameworks · RAG / AI features on existing data · Azure or GCP NOT A FIT FOR Data entry, scraping-only jobs, or work that needs Snowflake or dbt depth I don't have. I will say so if your post is one of those. U.S.-based (Chicago), veteran, respond within a few hours.