Use this data scientist job description as a free, editable starting point for your next post. It's built around decisions a data scientist actually owns, like whether an experiment worked or whether a model is good enough to ship, rather than a list of every library in Python. Below the template: the real day-to-day, U.S. wage data for the occupation, and first-round questions worth asking.
Data scientist job description template
What a data scientist does
A data scientist's work splits between analysis and modeling, and the mix depends heavily on the team. On a product team, most of the week may go to experiments: deciding what to measure, checking that randomization worked, and writing up whether a change moved the metric. On a team focused on machine learning, more time goes to features, model training, evaluation and working with engineers to deploy. In both cases, a large share of time is spent cleaning data and explaining results. O*NET's task list for the occupation reflects this mix: processing large datasets with statistical software, testing and validating models, comparing them on performance metrics, and presenting results to management.
Data scientists work with data engineers who build pipelines, analysts who handle reporting, ML engineers who run models in production, and the business owners who act on the answers. The most valuable ones turn a vague request into a question the data can actually answer.
At a startup, a data scientist is often the first data hire and does everything: sets up tracking, builds dashboards, runs the first experiments and fits a simple model. At a large company, the role is usually split into tracks, such as product analytics, machine learning or research, with more specialized tools and peer review of methods. Say which track you're hiring for, because the candidate pools barely overlap.
Skills and qualifications
Education norms are higher here than for most data roles. In O*NET's survey for Data Scientists, 48% of respondents said new hires need a bachelor's degree, 44% said a master's and 4% said an associate's degree. No license or certification is required. Requiring a PhD is rarely necessary outside research roles.
Salary
Data scientist is one of the few modern data titles with its own O*NET occupation: Data Scientists (15-2051.00). Using 2025 Bureau of Labor Statistics wage data, O*NET reports a national median wage of $120,230 per year, or $57.80 per hour. Pay at the 10th percentile is $67,240, and at the 90th percentile it reaches $199,130.
That wide range covers junior analysts with a data scientist title through to senior ML specialists, so place your role within it by level. In California, O*NET's median for the occupation is $141,590. Publish your range in the ad: a growing number of states and cities require pay ranges in postings, and data scientists tend to compare offers carefully.
Interview questions for data scientist candidates
- "An A/B test shows a 2% lift with p = 0.04. The product lead wants to ship. What do you check first?" Strong answers check sample ratio mismatch, whether the metric was pre-registered, novelty effects, multiple comparisons and whether 2% matters to the business.
- "Your churn model has 95% accuracy. Is it good?" Good candidates ask about the base rate. If only 5% of customers churn, predicting "no churn" for everyone scores 95%, so precision, recall and calibration matter more.
- "When would you choose logistic regression over a gradient-boosted model?" Listen for interpretability, small data, the need to explain decisions and ease of maintenance, balanced against accuracy.
- "Tell me about a model or analysis that didn't get used." The best STAR answers own the miss and explain what they changed about how they scope work.
- "Explain overfitting to a sales director." Look for a plain, accurate analogy and no jargon.
See our data scientist interview questions for the full set with model answers and a 1 to 4 rubric. For ML-heavy roles, add questions from our machine learning engineer interview questions.
Writing tips for this role
- Name the track. "Product data scientist (experiments and metrics)" and "ML data scientist (models in production)" attract different people. A post that asks for both usually gets neither.
- Don't require a PhD unless the work is research. Most applied roles need strong statistics and judgment, which many people with a bachelor's or master's degree have.
- Skip the buzzword stack. Listing TensorFlow, PyTorch, Spark, Hadoop and every cloud signals confusion. Name what the team uses today.
- Describe a real problem they'll solve. "Forecast weekly demand for 2,000 SKUs" is more attractive than "drive insights across the business."
- Check you need a data scientist at all. If the job is mostly dashboards and reporting, our data analyst job description may be the better fit, and you'll reach a larger pool.
Screening data scientist candidates with AI
Data science applicants often look alike on paper, with the same libraries and course projects. NTRVSTA ranks each resume against your job description, then an AI interviewer conducts a live first round, by video avatar or audio only, at any time that suits the candidate. The interview adapts in real time: when someone explains the base-rate problem well, the next question goes deeper. Python and AI/ML topics are in the exam library, and you can add your own experiment and modeling scenarios per role.
Your hiring manager reviews a score breakdown, video snippets of key moments and a shareable report before committing team time to the case round. To understand what drives those scores, read how AI interviews are scored, and use the interview scorecard template for the human rounds.
Questions
What's the difference between a data scientist and a data analyst?
A data analyst mostly answers questions with existing data, through queries, dashboards and reports. A data scientist adds statistical modeling, experiment design and prediction, and is more often expected to write production-quality Python. Many companies blur the line, so describe the actual work rather than relying on the title.
Should a data scientist job description include a take-home assignment?
Only if it's short, around two to three hours, uses a small realistic dataset and comes with clear scoring criteria. A live case conversation often works just as well and doesn't ask candidates with jobs to give up a weekend.
Do data scientists need machine learning engineering skills?
It depends on your team. If data scientists deploy their own models, list deployment and monitoring as must-haves. If ML engineers handle production, list it as nice to have and focus the post on modeling and analysis.
See how an adaptive AI interviewer handles a data science first round: book a 20-minute demo, or start a sample interview yourself. More templates live in our job descriptions library.
