A good data scientist spends less time on algorithms than most job ads suggest. The harder part of the job is turning a vague business worry, such as “why are customers leaving?” or “are we pricing this right?”, into a question the data can actually answer, then building a model that answers it and explaining the result to the people who will act on it. A job description that only lists Python libraries attracts people who are good at the middle step and weak at the other two.
This page gives you a complete data scientist job description you can copy and adapt, followed by guidance on how the role changes with seniority, which skills to test during screening, the mistakes that make these ads attract the wrong applicants, and how to choose between contract and permanent hiring.
Data scientist job description template
Replace the text in square brackets and remove anything that does not apply.
About the role
[Company] is looking for a Data Scientist to join our [team name] team in [city / remote]. You will work with product, business and engineering colleagues to frame important questions, build and validate models for problems such as [forecasting / churn / pricing / risk], and explain what the results mean in plain language. You will report to [manager title] and work mainly with data from [main data sources].
Responsibilities
- Work with business stakeholders to turn open-ended problems into clear, testable hypotheses and agreed success measures.
- Explore, clean and join data from [warehouse / lakehouse / source systems] using SQL and Python.
- Build, validate and compare models for [forecasting, churn, pricing, segmentation or risk], choosing the simplest approach that meets the need.
- Design and analyze A/B tests and other experiments, including sample size, guardrail metrics and interpretation of results.
- Check models for leakage, bias and drift, and document assumptions and known limits.
- Present findings to non-technical audiences in a way that leads to a decision, with clear charts and written summaries.
- Work with data engineers to make sure the features and data your models need are reliable and available in production.
- Hand over models to [ML engineering / platform team] or deploy them yourself, and monitor their performance after launch.
- Keep code, notebooks and analyses reviewable and reproducible through version control and shared standards.
Requirements
- [X]+ years of experience in a data science, analytics or applied statistics role.
- Strong Python, including pandas, NumPy and scikit-learn, and confident SQL on large tables.
- Solid grounding in statistics: probability, hypothesis testing, regression and experimental design.
- Hands-on experience building and validating supervised models, including gradient boosting methods such as XGBoost.
- Experience designing or analyzing A/B tests in a business setting.
- Ability to explain technical results clearly in writing and in meetings to people without a data background.
- A degree in [statistics, mathematics, computer science, economics or a related field], or equivalent practical experience.
Nice to have
- Experience with Apache Spark or Databricks for working with large datasets.
- Familiarity with time series forecasting or causal inference methods.
- Experience building dashboards in Power BI or Tableau to share results.
- Domain knowledge of [your industry, such as retail, lending or logistics].
What we offer
- [Salary range] and [bonus or equity details].
- [Work mode: on-site / hybrid / remote] from [location].
- [Learning budget, conference time or certification support].
How the data scientist role changes by level
The title stays the same across levels, but what you are paying for changes. Write the level into the ad so applicants can judge their fit.
Junior data scientist. A junior person works on well-defined problems that someone else has framed. Their responsibilities should center on data preparation, exploratory analysis, building baseline models and producing clear reports under review. Requirements can lean on fundamentals: statistics coursework, Python and SQL, and one or two complete projects they can walk through, whether from work, study or their own time.
Mid-level data scientist. At this level the person owns a problem from question to result. They should be able to sit with a product manager, agree what success looks like, choose a method, run the experiment or build the model and present the outcome without close supervision. Ask for evidence of models or experiments that changed a real decision, not just models that were built.
Senior or lead data scientist. A senior hire decides which problems are worth solving at all. The job description should mention prioritizing work with leadership, reviewing other people’s methods, setting standards for validation and experimentation, and mentoring. The stronger signal is judgment: knowing when a simple rule or a good report will beat a model.
Data scientist skills that matter most, and how to check them
Most CVs in this field list the same tools. The screening process has to separate people who have used them on messy business problems from people who have followed tutorials.
- Problem framing. Give the candidate a short business scenario, such as a subscription business that thinks churn is rising, and ask how they would approach it before touching any data. Listen for questions about how churn is defined, what decision the analysis will support, and what data exists.
- Statistics and experimentation. Ask them to explain how they would size an A/B test and what they would do if the test result was positive but a guardrail metric got worse. Strong candidates talk about practical significance, not only p-values.
- SQL. A short live or take-home exercise with joins, window functions and a deliberately awkward edge case, such as duplicate rows or missing dates, tells you more than any certificate.
- Modeling and validation. Ask them to walk through a model they built: how they split the data, what baseline they compared against, and how they checked for leakage. Real practitioners can describe what went wrong.
- Communication. Ask them to explain a past result as if you were the head of sales. Jargon here predicts trouble in your weekly review.
A small take-home task built on an anonymized slice of your own data works well, as long as it is scoped to a few hours and you review it with the candidate rather than just scoring it.
Common data scientist job description mistakes
- Listing every tool in the field. An ad that asks for deep learning, NLP, computer vision, Spark, Kubernetes and Tableau describes several jobs. Fix: name the problems the person will work on and the tools your team actually uses.
- Hiring a data scientist when you need a data engineer. If your data is not yet centralized or reliable, a data scientist will spend most of their time building pipelines. Fix: be honest about the state of your data in the ad, or hire a data engineer first.
- Leaving out who the person reports to. Data scientists who sit in a product team work very differently from those in a central analytics team. Fix: state the reporting line and the main stakeholders.
- Treating a PhD as the default. Some research-heavy roles need one, but most business data science does not. Fix: ask for the specific skills, and say “or equivalent practical experience”.
- No mention of how models reach production. Candidates want to know whether their work will be used. Fix: explain whether they deploy models themselves or hand them to an engineering team.
Contract, contract-to-hire or permanent: which suits a data scientist
A contract data scientist makes sense for a defined piece of work with a clear end, such as building a demand forecast for a new region, running an analysis ahead of a pricing change, or validating a model before a regulatory review. You get the skill for the length of the project without adding to permanent headcount.
Contract-to-hire suits teams that are building their first data science capability and are not yet sure what profile they need. A few months of real work shows whether the person frames problems well and fits with your stakeholders.
Permanent hiring is the right call once data science is part of how the company makes decisions every week, because the value of the role grows as the person learns your data, your customers and your history of past experiments. Zilo AI can help with any of the three: we hire data scientists in India on contract, contract-to-hire or permanent terms, screening candidates on statistics, SQL, Python and real project work before you meet them. If you are hiring across several technical roles at once, our IT hiring services cover the wider team.
Frequently asked questions
What should a data scientist job description include?
It should describe the business problems the person will work on, the data and tools they will use, who they will report to, and the level of independence expected. Add the core requirements in statistics, Python and SQL, a short list of nice-to-have skills, and practical details such as work mode and salary range.
What is the difference between a data scientist and a data analyst?
A data analyst mostly describes what has happened, through reports, dashboards and ad hoc queries. A data scientist builds models and experiments to predict what will happen or to test what causes an outcome. In smaller companies the roles overlap, so describe the actual work rather than relying on the title.
What is the difference between a data scientist and a data engineer?
From the data scientist’s side, the engineer is the person who makes sure the data arrives clean, on time and in the right shape. The data scientist then uses that data to answer questions and build models. A team with data scientists but no engineering support usually ends up with its scientists maintaining pipelines.
Should a data scientist job description ask for machine learning engineering skills?
Only if the person will deploy and run models in production without an ML engineering team. If you have such a team, ask for the ability to hand over clean, documented code instead, and keep the deployment skills in the nice-to-have list.
If you would rather spend your time interviewing a shortlist than writing and posting the ad, Zilo AI can find and screen data scientists for you. Get in touch with our team and tell us about the problems you want the role to solve.