went through the Accenture data engineer interview process earlier this year for a senior DE position on one of their cloud migration teams. posting this because the DE-specific experience is harder to find than the SWE or analyst content.
process was 3 rounds after the recruiter screen.
first technical round (60 min): heavy SQL. we covered CTEs, window functions, and a tricky problem involving recursive CTEs to traverse a hierarchy. I'll be honest, I had to think out loud on the recursive one. they were fine with that, seemed interested in how I reasoned through it. also got a question about database normalization: when do you denormalize and why. good old data warehouse tradeoffs.
second technical round (60 min): pipeline architecture. they gave me a scenario: a client gets a 500GB CSV dump every night and needs it queryable by 9am. walk me through the architecture. I talked through ingestion (landing zone in S3 or ADLS), transformation (Spark or dbt depending on client stack), orchestration (Airflow), monitoring, and SLAs. they probed on what happens when the file is late, what happens when it has schema drift, and how I'd alert the client team. real-world messy stuff. no LeetCode algorithms.
behavioral round (45 min): lots of delivery-focused questions. "tell me about a pipeline that failed in production and how you handled it." "describe how you managed expectations with a non-technical stakeholder when a project slipped." Accenture cares a lot about client-facing skills even for technical roles. if you have zero client communication experience, at least prepare stories that show you explained technical things to non-engineers.
no take-home. no system design whiteboard in the FAANG sense. the architecture discussion replaced that.
the tech stack on the job is pretty fluid: Azure, GCP, Snowflake, Databricks depending on whatever the client uses. so they're evaluating cloud-agnostic reasoning rather than specific tool mastery.
offer for senior DE came in at ~$130k base in a mid-tier city. consulting firms tend to have narrower bands than big tech but the project variety is real.