I sat in my 300-level database concepts class and wrote SELECT * FROM users JOIN orders ON users.id = orders.user_id. I got an A. I felt like a genius. I genuinely believed I was ready for a Data Engineering role at a top-tier tech company.
Then I got my first technical screen with a MAANG company.
The prompt wasn't about joining two tables. It was about calculating a 7-day rolling average of user retention, partitioned by cohort, and filling in gaps for days where users had zero activity.
I bombed it. Hard.
That was the day I realized there is a massive, gaping hole between academic SQL and what the industry actually demands.
The Problem with College Databases
Academic curriculums are great at teaching you the theory of normalization (Boyce-Codd, anyone?) and the absolute basics of CRUD operations. But they almost universally skip the analytical functions that modern data teams use every single day.
If you are interviewing for a Data Analyst, Data Engineer, or Backend role today, you aren't going to be tested on 1st Normal Form. You are going to be tested on:
Advanced Window Functions (ROW_NUMBER(), RANK(), LEAD(), LAG())
Recursive CTEs (Common Table Expressions)
Self-Joins (Comparing rows within the same table)
Date Math & Cohort Analysis
The Headache of Practicing
After I bombed that interview, I knew what I needed to study. But I quickly realized that practicing advanced SQL is an absolute nightmare.
You essentially have two options:
Pay $200 for a bloated video course. You watch someone else write code on a whiteboard for 10 hours. You don't actually write any code yourself, so you forget the syntax the second the video ends.
Set up a local environment. You spend your entire Saturday downloading Docker, installing Postgres locally, setting up a database client like DBeaver, and hunting down CSVs of fake data on Kaggle just so you can practice writing a single query.
It’s exhausting. The friction to just "write some code" is incredibly high.
Building dataDrill
I got so frustrated with the process that I decided to build the tool I wished I had when I was studying for that first interview.
I built dataDrill.
It is an entirely in-browser practice platform. You don't install Docker. You don't configure Postgres. You don't download CSVs.
You just open the website, and you instantly have a real PostgreSQL engine (running locally in your browser via WASM) with tables already populated with data.
I sourced 90+ real questions directly from technical screens at companies like Meta, Google, Stripe, and Netflix. I specifically curated the questions to focus on the exact things college skipped—the window functions, the self-joins, the date math.
I also added line-by-line solutions so that when you inevitably get stuck on a recursive CTE, you can see exactly why a specific piece of syntax is used, rather than just copying and pasting a final answer.
If you are a CS major about to graduate, or a junior dev trying to break into Data Engineering, don't walk into your interview thinking GROUP BY is enough.
I've left several of the hard challenges on the site completely free. Try them out, test your skills, and bridge the gap before your interviewer exposes it.
Top comments (0)