Interview scorecard templates

Data Scientist interview scorecard template

Score data science candidates on rigor and on whether their work would change a decision. This scorecard covers statistics, modeling, experimentation, programming and business understanding, with questions to ask and scored examples for each.

Customize this scorecard

Free to use. Nothing you type on the scorecard is saved.

Interview scorecard

Data Scientist

Score each skill from 1 to 5. A 1 is below the bar for this role, a 3 meets it, and a 5 is above it. The examples show what each sounds like. Use 2 and 4 for answers that fall in between.

Statistics

Applies statistical reasoning to draw sound conclusions from messy data and explains uncertainty plainly.

Score for Statistics

Questions

  • Tell me about a time statistics changed a conclusion people were ready to accept.
  • Explain a confidence interval to a manager who has never studied statistics.
  • Tell me about a time your data was biased or incomplete. What did you do?

1: Below the bar

Quotes methods or p-values without explaining them. Treats a single number as certain and ignores bias in the data.

3: Meets the bar

Explains core ideas correctly in plain words. Notices common problems like small samples or bias, sometimes only when asked.

5: Above the bar

Explains uncertainty simply with a real example. Spots data problems unprompted and says how they would change the conclusion.

Modeling

Builds predictive models suited to the problem and checks that they work on new data.

Score for Modeling

Questions

  • Tell me about a model you built. Why did you choose that approach?
  • Tell me about a time a model performed worse than expected. What did you do?
  • How did you decide your model was good enough to use?

1: Below the bar

Picks a method because it is familiar or popular. Cannot say how they checked it works on new data.

3: Meets the bar

Explains why the approach fits the problem and tests on held-out data. Handles problems with some guidance.

5: Above the bar

Compares simple and complex options, explains the choice, and finds why a model failed. Links the quality measure to its real use.

Experimentation

Designs and interprets tests that show whether a change caused a measurable difference in results.

Score for Experimentation

Questions

  • Tell me about an experiment you designed. How did you decide what to measure?
  • Tell me about a test where the result was unclear or surprising. What did you do?
  • Describe a time you could not run a proper test. How did you still reach an answer?

1: Below the bar

Compares before and after without a control group. Takes any difference as proof and cannot say what else could explain it.

3: Meets the bar

Sets up a control group and a clear success measure. Explains results correctly but may miss sample size or side effects.

5: Above the bar

Defines the measure and test size upfront, checks for flaws, and explains what to do when a test is impossible.

Programming

Writes clear, working code that others can read, rerun and trust to analyze data.

Score for Programming

Questions

  • Tell me about a piece of analysis code you wrote that others had to reuse.
  • Describe a time your code gave a wrong result. How did you find the cause?
  • How do you check that your data cleaning code does what you intended?

1: Below the bar

Code works only on their own machine. Can't describe how they test it or find bugs.

3: Meets the bar

Writes readable code others can rerun. Checks results by hand and can describe how they tracked down a past bug.

5: Above the bar

Writes tested, reusable code with clear structure. Checks data at each step and explains how they caught and prevented bugs.

Business understanding

Connects data work to company goals, customers and money, and shapes analysis around real decisions.

Score for Business understanding

Questions

  • Tell me about a time you changed an analysis after learning what the business needed.
  • Describe a project where you chose what to work on based on its value. How?
  • Tell me about explaining a result to someone who had to make a decision.

1: Below the bar

Talks only about methods and data. Can't say who uses the results or what decision they support.

3: Meets the bar

Names the decision and the stakeholder behind a project. Links results to a business goal but gives vague impact.

5: Above the bar

Starts from the decision, weighs value against effort, and shows measured business impact. Adjusts the message to the audience.

Overall recommendation

Decide on your own, before you compare notes with the rest of the panel.

Overall recommendation

Use this scorecard in PontaHR

With PontaHR, your team fills out scorecards on the candidate's profile. Compare ratings and feedback side by side when it's time to decide.

See how scorecards work in PontaHR, then try it with a guided free trial.

More interview scorecard templates

Hiring for a different role? Make your own scorecard