SQL practice that looks like the job.

Why we built this

I’m Anthony. I’ve been the new analyst who got handed a messy database and a stakeholder who didn’t know what they wanted — and I’ve also been the senior who watched new hires struggle with the same thing year after year. Every SQL course I tried taught syntax in isolation. The first week on an actual job was a completely different skill, and nobody was teaching it.

Knowing SELECT from JOINisn’t the hard part. The hard part is translating a vague business question into a query that produces a trustworthy answer from imperfect data, on a deadline, for a stakeholder who doesn’t care about your CTE. That gap — between textbook SQL and job SQL — is what caseSQL is built to close.

So every mission opens with a Slack message from a “manager” with a vague ask. The schemas are intentionally messy — NULLs, duplicates, refunds that arrive after the sale, timestamps in three timezones. The validator doesn’t say “right” or “wrong”; it responds the way a senior analyst reviewing your PR would. The point isn’t to ace an interview. It’s to make your first 90 days on the job feel like a month you’ve already lived.

caseSQL is solo-built and bootstrapped. No VC pressure to shoehorn in AI features nobody asked for, no retention-metric optimizations that make learning worse. Pro exists so this can keep being my full attention; the free tier stays genuinely useful forever. If you ever want to tell me what’s broken or what’s missing, the feedback form is one click from every page and reaches my inbox directly.