Photo by the Author
The obvious Getting started
Everyone knows what comes up in data science discussions: SQL, Python, machine learning models, statistics, sometimes program design or case studies. If this comes up in discussions, that's what they're testing, right? Not really. I mean, they are convinced of everything I wrote, but they don't stop at that only: There is a hidden layer behind all those technical activities that companies are actually testing.
Photo by author | Imgflip
It's probably confusing: while you think you're showing off your coding skills, employers are looking for something else.
That other thing is the hidden curriculum – the skills that will determine whether you can succeed in the role and in the company.
Photo by author | Napkin ai
The obvious 1. Can you translate business to data (and back)?
This is one of the biggest skills needed for data scientists. Employers want to see if you can take a clear business problem (eg “Which customers are the most important?”), turn it into a data analysis or machine learning model, and distill the information back into a decision language.
What to expect:
- The case study is loosely constructed: for example, “our active daily users are flat. How can you improve engagement?”
- The following questions force you to confirm your analysis: for example, “Which metric can you track to know if sales are improving?”, “How does management care about revenue,
What exactly does it test:
Photo by author | Napkin ai
- Clarity: Can you explain your points in clear English without too many technical terms?
- Prioritization: Can you highlight the key insights and explain why they are important?
- Audience awareness: Do you change your language according to your hands (technical vs. ang-technical)?
- Confidence without being arrogant: Can you clearly explain how you can explain clearly, without being overly defensive?
The obvious 2. Do you understand trading?
In your work, you will have to make a trade-off, accuracy VS. Interpreting or changing bias vs. to be different. Employers want to see you do that in interviews, too.
What to expect:
- Questions like: “Would you use random forest or systematic regression here?”.
- There is no right answer: Situations where both answers can be right, but they are interested in why you choose.
What exactly does it test:
Photo by author | Napkin ai
- There is no national “best” model: Do you understand that?
- Sales transplant – can you do that with plain words?
- Business Alignment: Do you demonstrate an awareness of aligning your model choices with business needs, instead of chasing technical perfection?
The obvious 3. Can you work with incomplete data?
Details in conversations are rarely clean. There are missing standards, duplicates, and other inconsistencies. That's on purpose to show the actual data you'll have to work with.
What to expect:
- Incomplete Data: Tables with inconsistent formats (eg
- Analytical reasoning question: Questions about how to validate an assumption
What exactly does it test:
Photo by author | Napkin ai
- Your nature of data quality: Do you pause and question details instead of mindlessly coding?
- Prioritizing data cleaning: Do you know which issues should be cleaned first and have the greatest impact on your analysis?
- Judgment under ambiguity: Do you make the reasoning clear so that your analyst is clear and you can move forward while acknowledging the risk?
The obvious 4. Are you thinking about exams?
Experimentation is a big part of data science. Even if the role doesn't explicitly test, you'll have to do A/B tests, pilots, and validation.
What to expect:
What exactly does it test:
Photo by author | Napkin ai
- Your test design skills: Do you correctly define control vs. Treatment, randomization, and sample size?
- Critical interpretation of results: Do you consider statistical significance vs. Functional significance, confidence intervals, and secondary effects when interpreting test results?
The obvious 5. Can you stay calm under ambiguity?
Most conversations are designed to be confusing. Interviewers want to see how you work with incomplete and incomplete information and instructions. Notice what, that is exactly what you will find in your own work.
What to expect:
- Ambiguous questions with missing form: for example, “How would you measure customer engagement?”
- Going back to your clarifying questions: for example, you can try to clarify the above by asking, “Do we want engagement measured in time spent or number of sessions?”. Then the interviewer can put it when he asks, “What would you choose if the leadership came?”
What exactly does it test:
Photo by author | Napkin ai
- Mindset under Uncertainty: Are you Cold, or Calm and Pragmatic?
- Problem structure: Can you place an order with a vague request?
- Making assumptions
- Business thinking: Do you tie your thoughts to business goals or some counterintuitive assumptions?
The obvious 6. Do you know that “better” is the enemy of “good”?
Employers want you to be pragmatic, which means: Can you provide useful results as soon as possible? A candidate who would spend six months improving the model's accuracy by 1% is not exactly what they want, to put it mildly.
What to expect:
- Pragmatism Question: Can you come up with a simple solution that solves 80% of the problem?
- Assessment: The interviewer presses you to explain why you are going to quit.
What exactly does it test:
Photo by author | Napkin ai
- Verdict: Do you know when to stop doing well?
- Business alignment: Can you connect solutions for business impact?
- Resource awareness: Are you respectful of time, cost, and team capacity?
- Smart Iterative: Do you ship something useful now, then improve over time, instead of spending more time creating the “perfect” solution?
The obvious 7. Can you handle pushback?
Data science is collaborative, and your ideas will be challenged, so the discussions have multiplied that.
What to expect:
- Critical Thinking Test: Interviewers try to stimulate you and throw poke holes your way
- Alignment test: Questions like, “What if leadership disagrees?”
What exactly does it test:
Photo by author | Napkin ai
- Strength under sclitiny: Do you stay calm when your path is a challenge?
- Clarity of thinking: Are your thoughts clear to you, and can you explain them to others?
- Flexibility: When an interviewer reveals a pitfall in your approach, how do you react? Do you agree graciously, or do you stumble and run out of the office crying and complaining and propose?
The obvious Lasting
You see, technical discussions aren't really what you thought they were. Remember that all technical tests are actually:
- Interpreting Business Problems
- Managing Trade-offs
- Handling messy, idiosyncratic and conditional data
- Knowing when to act and when to stop
- Working together under pressure
Nate receipt He is a data scientist and product strategist. He is also a self-proclaimed educationalist, and the founder of Stratascratch, a platform that helps data scientists prepare for their interviews with real interview questions from top companies. Nate writes on the latest trends in the job market, gives interactive advice, shares data science projects, and covers all things SQL.
