Intuit · Primly Community

Just finished my Intuit data engineering loop. here's what actually happened

de_derek · 4 replies

went through the full process for a senior data engineer role on the QuickBooks data platform team. figured i'd write it up while it's fresh.

Round breakdown: recruiter screen: 30 min, pretty standard. talked about scale, my pipeline experience, rough comp range. recruiter was sharp and moved fast. technical phone screen: CoderPad, 45 min. one medium SQL problem (window functions, find customers with consecutive months of inactivity), one Python/pandas question. not hard, but they time you on the SQL. virtual onsite: 4 panels over one day data modeling: design a transaction enrichment pipeline for financial data. latency vs. cost tradeoffs, partitioning strategy. coding: graph traversal problem. straightforward if you've done any LC mediums. behavioral x2: both interviewers asked about customer-obsession stories, one asked about a time you dealt with bad data in production and the downstream impact.

What surprised me: they asked very specifically about how you handle data quality issues that could affect a customer's tax filing. not generic "tell me about a data quality problem". the financial domain specificity was real.

What I'd prep: know your SLAs, know what downstream consumers your pipelines fed, have a story about catching something before it caused a customer-facing problem.

Got an offer. total process was 3.5 weeks from recruiter reach-out to offer call.

4 replies

ds_dmitri

the window function SQL question is interesting. was it a rolling window or a session-based gap analysis? asking because I've seen both at fintech companies and they feel very different to write under time pressure.

de_derek

gap analysis. you had a table of login events and had to find users who hadn't logged in for 90 consecutive days. the trick was the self-join vs LAG() tradeoff. i used LAG() and it was fine.

backend_bekah

the financial domain specificity tracks. i did a backend loop for a TurboTax team and the design round was "design the calculation engine for estimated quarterly tax reminders." not a generic design question at all. you need to at least have thought about what it means when you get a calculation wrong.

staff_steph

3.5 weeks is genuinely fast for a company this size. they've clearly cleaned up their process. i remember Intuit being notoriously slow a few years back.