Cardinal Health Data Engineer Interview Guide

Offers get lost when candidates treat the compliance layer as a checkbox rather than a design constraint. Interviewers here have seen pipelines fail audits because an engineer built a technically correct solution that didn't preserve row-level lineage or dropped rejected EDI records instead of quarantining them. If you walk through a pipeline design and your error-handling strategy is "log and skip", expect a follow-up about what happens when that skipped record is a controlled-substance transaction. The inverse behavior that reads as a hire is someone who proactively asks about data retention policy and audit requirements before proposing a schema. The other failure mode is overclaiming on tooling: candidates who frame their prep around Kafka or Flink and pitch real-time rewrites of batch processes tend to read as misaligned with how this engineering organization actually runs.

Cardinal Health's loop is designed to find engineers who can own pipelines where a data quality failure has regulatory consequences, not just an angry stakeholder. The company's distribution network handles controlled substances, specialty pharmaceuticals, and nuclear pharmacy products, and the data infrastructure has to satisfy DEA traceability requirements and FDA compliance audits alongside normal SLA obligations. What the interviewers are filtering for is a particular kind of rigor: candidates who instinctively think about auditability, schema stability, and failure recovery before they think about throughput optimization. They also need people who can operate without much tooling autonomy, because Cardinal Health's technology choices move slowly and by committee. If your instinct when a pipeline breaks is to ask "what does the compliance trail look like" before "how fast can I restart it", that instinct will read well here.

Start prep with data modeling for regulated domains: practice SCD designs where you also have to produce a lineage report, and think through how you'd handle rejected or quarantined records in a warehouse load. That earns more than rehearsing streaming problems this loop is unlikely to surface. EDI familiarity is worth at least a few hours if you don't have it; you don't need to implement a parser, but you should be able to discuss what an 856 advance ship notice contains and how you'd load it into a staging table. At the L6 level, interviewers will push on how you handle scope ambiguity in compliance-adjacent work: when a business rule is unclear, do you escalate or make an assumption? Have a concrete answer ready. Given that 10 salary reports exist for this loop, and the ladder compresses into 2 levels, leveling conversations are simpler than at larger engineering orgs, but you should still anchor your expected offer to $144K and be prepared to make the case for staff scope.