Senior AI / Data Engineer: Medicare member eligibility module (Databricks)

We run member and prospect outreach for Medicare Advantage clients. The full system already works: ingest, identity, scoring, audience build, vendor file, matchback. We are not rebuilding it. One module is the problem. Eligibility, suppression, and release checks still depend on a large Excel workbook. The Databricks job that consumes those rules is in production. After a recent Snowflake permission change it has been failing, and even when it succeeds we do not fully trust the file. A wrong include or a missed suppression is a compliance issue, not a metric miss. PHI leaves the building. We need a senior engineer who can treat this as a system repair, not a notebook exercise. Read the live PySpark / SQL job. Trace the workbook into executable rules. Use an LLM only to accelerate extraction and to draft QC, then verify every rule against source tables before anything is trusted. Write the checks that catch duplicates, bad exclusions, test/control leakage, and count breaks before release. This is data engineering with model-assisted rule work. If your background is only model fitting, this is not the right job. If you have shipped person-level pipelines and can keep an LLM from inventing a business rule, it is. Stack on this module: Databricks, PySpark, Snowflake, SQL, Python. Git, AWS, SFTP, and Linux show up around the job. Excel is the current spec. US-based only. Background check, OIG screen, HIPAA training, and a BAA are required before data access.