Just wrapped up an Apple data engineer interview loop last month, E4 level targeting the AppleID / IAM data platform team. Sharing what I saw because there was basically nothing useful online when I was prepping.
The loop was five rounds total: one recruiter screen, one technical phone screen, then three onsite rounds (all virtual in 2026 but they still call it onsite).
Technical phone screen. One hour with a staff DE. Half SQL, half system design lite. The SQL question was a multi-step thing: find the top 3 devices per country by active sessions in the last 7 days, with some nuance around deduplication. Window functions, CTEs. Not trivial but also not trying to trick you. The other half was a conversation about a pipeline I'd built, how I handled late-arriving data, what our SLA was.
Onsite rounds. Round 1 was pure SQL and data modeling. Given a schema (events table, user table, device table) and asked to answer 4 progressively harder questions. Last one involved calculating a rolling retention metric. Know your window functions cold. Round 2 was pipeline architecture. "Design a real-time anomaly detection pipeline for App Store purchase events at Apple's scale." They cared a lot about exactly-once semantics, what happens when a Kafka consumer dies mid-batch, how you'd handle schema evolution. I talked Flink/Spark Structured Streaming; they seemed fine with either. Round 3 was behavioral. Standard stuff but specifically Apple-flavored: they pushed hard on "owning the outcome" and cross-functional collaboration. Think times you pushed back on product or convinced a skeptical stakeholder. They really probe for autonomy.
Things I didn't expect. Almost no Spark-specific trivia. More emphasis on data quality and pipeline observability than raw throughput. They asked about how I'd monitor a pipeline and what metrics I'd alert on.
Total timeline was about 7 weeks from first recruiter contact to offer. The debrief took almost two weeks which was nerve-wracking.
Happy to answer specifics if you're interviewing for a DE role there.