Free Guide

Atomic Data
Interview_

40 real interview questions asked at data science companies — plus the Atomic Habits system that gets you practicing them consistently.

SQL Statistics Machine Learning Python

Free forever · No spam · Unsubscribe anytime

40
Questions
4
Domains
90
Day plan
$0
Total cost
The Problem

You're reviewing.
Nothing's sticking.

You're not underprepared. You're practicing wrong. Most interview prep creates an illusion of progress — watching solutions, reading explanations, recognising answers. Recognition isn't retrieval. And retrieval is what interviews test.

It's not a motivation problem. It's not a knowledge problem.
It's a systems problem — and systems are fixable.
About the Author

Written by someone
who's done it.

Pénélope Lafeuille
Pénélope Lafeuille
@penelope_data · Data Scientist · Builder

I failed a SQL screen on a question I'd seen in a tutorial — one I'd watched twice. The problem wasn't knowledge. I recognised the concept immediately after getting it wrong. The problem was I had never once tried to answer it without the video in front of me.

That failure sent me back to first principles. I rebuilt my prep system using James Clear's Atomic Habits — not as a motivation hack, but as a structural framework. Tiny daily reps. Domain rotation. Weekly simulation.

This guide is the system I wish I'd had. It's free because it should be.

What You Get

A system, not a syllabus.

40 questions across the four domains every data science interviewer tests. Each with structured answers and a daily practice system to build real retrieval — not passive recognition.

DOMAIN 01
SQL
Joins, window functions, aggregations, subqueries — the questions that trip up candidates who only practiced in guided environments.
DOMAIN 02
Statistics
Hypothesis testing, p-values, distributions, A/B testing — explained so you can say it out loud, not just recognise the answer.
DOMAIN 03
Machine Learning
Bias-variance, regularisation, evaluation metrics, model selection — the conceptual layer interviewers probe for real depth.
DOMAIN 04
Python
Pandas, NumPy, data manipulation, writing clean functions — practical data science Python under realistic time pressure.
section_01 The Problem with Passive Review mindset
section_02 The Retrieval Practice Framework method
section_03 SQL — 10 Questions SQL
section_04 Statistics — 10 Questions statistics
section_05 Machine Learning — 10 Questions ML
section_06 Python — 10 Questions python
section_07 The Daily Practice System habits
section_08 Domain Rotation Schedule schedule
section_09 The Weekly Simulation simulation
section_10 Your 90-Day Plan action
Framework Source

Atomic habits meet
technical interview prep.

This guide applies James Clear's four laws of behaviour change to the specific problem of data science interview preparation. One rep per day. Domain rotation. Weekly simulation.

James Clear — Atomic Habits
framework source
James Clear — Atomic Habits

The #1 New York Times bestseller on building systems that make good behaviours inevitable. This guide applies that exact framework — the four laws, habit stacking, identity-based habits — to the specific problem of data science interview preparation.

How It Works

One rep a day.
Compounded over 90.

STEP 01
One rep per day
The daily minimum: one question, answered out loud, without looking. Small enough that skipping feels worse than doing it.
STEP 02
Rotate domains
SQL on Monday. Statistics on Tuesday. ML on Wednesday. Python on Thursday. Built-in rotation prevents burnout and builds breadth.
STEP 03
Simulate weekly
Every Sunday: one random question, three-minute timer, no notes. Retrieve → identify gap → fill gap → retrieve again.
STEP 04
Review and adjust
A weekly four-question review in writing. What stuck. What slipped. What to shift toward. Compounds across 90 days.
Is This For You?

This guide has a
specific reader in mind.

→ This is for you if...
  • → You're a data science student or career switcher preparing for your first or next role
  • → You have SQL and Python exposure but haven't practiced answering questions out loud
  • → You've tried study plans before and dropped them — and want a system small enough to stick
  • → You know the concepts but freeze when someone asks you to explain them
✗ This is not for you if...
  • ✗ You're a senior data scientist with 5+ years looking for advanced material
  • ✗ You want deep learning or research-level ML — this covers practical generalist roles
  • ✗ You're looking for a one-time cram session the night before an interview
  • ✗ You already have a daily practice system that's working

Your first rep
is one email away.

Enter your email and I'll send you the Atomic Data Interview guide — free, immediately, no strings.

Free forever · No spam · Unsubscribe anytime