Data scientist job description template (2026)
A complete data scientist job description you can copy, plus seniority levels and tips for hiring someone whose analysis actually changes decisions.
By the Ryz Labs team · Updated October 2026
A data scientist job description should name the decisions the person will inform. "Data scientist" can mean an experimentation lead, a forecasting specialist, a product analyst with Python, or someone who builds and deploys models. Say which questions the business needs answered, how results reach decision makers, and whether the role ships production models or hands them to engineers. The template below is written for a senior data scientist focused on experimentation, causal analysis and statistical modeling that shapes product and business decisions. If you need someone to train and serve models in production, a machine learning engineer description is usually a better fit.
Data scientist job description template
Job title
Senior Data Scientist (Experimentation and Decision Science)
Employment type: full-time or contract. Location: remote, with at least four hours of overlap with US Eastern time.
About the role
We are looking for a senior data scientist to help [company name] make better decisions with data. You will design and analyze experiments, build models that forecast and explain [demand / churn / conversion / risk], and work directly with product, marketing and leadership to decide what to build and where to invest. Our data lives in [Snowflake / BigQuery / Databricks] and is modeled in [dbt]. You will report to [title] and partner with data engineers who own pipelines.
Responsibilities
- Design A/B and multivariate tests with clear hypotheses, primary metrics, guardrail metrics, minimum detectable effects and sample size calculations.
- Analyze experiments correctly, including variance reduction with CUPED, sample ratio mismatch checks, and handling of novelty effects and peeking.
- Use causal inference methods such as difference-in-differences, synthetic control, regression discontinuity or instrumental variables when randomized tests are not possible.
- Build forecasting models for demand, revenue or capacity, with honest uncertainty intervals and backtests.
- Build predictive models for churn, conversion or risk, and explain their drivers so teams can act on them.
- Define and maintain core metrics with product and data engineering, including precise definitions and known caveats.
- Turn ambiguous business questions into analyses with a clear recommendation, and say what the data cannot answer.
- Present results to non-technical stakeholders in short written memos and charts, and defend the method under questioning.
- Improve experimentation tooling and standards, such as templates, power calculators and review checklists.
- Review other analysts' and scientists' work for statistical errors and misleading charts.
Requirements
- 5+ years in data science, analytics or quantitative research, with at least 3 years influencing product or business decisions.
- Strong foundations in statistics: hypothesis testing, confidence intervals, regression, power analysis and multiple comparisons.
- Hands-on experience designing and analyzing online experiments, including diagnosing invalid tests.
- Working knowledge of causal inference methods for observational data and their assumptions.
- Advanced SQL for pulling and validating your own data from a warehouse.
- Strong Python or R with pandas or Polars, statsmodels, scikit-learn or equivalent libraries.
- Experience with at least one modeling family beyond linear regression, such as gradient boosted trees, survival models or Bayesian hierarchical models.
- Excellent written communication: you can explain a confidence interval to an executive without hiding the uncertainty.
- A degree in a quantitative field or equivalent practical experience.
Nice to have
- Bayesian methods with PyMC, Stan or NumPyro.
- Experience with experimentation platforms such as Eppo, Statsig, GrowthBook or an in-house system.
- Marketing mix modeling or multi-touch attribution.
- Time series libraries such as statsforecast, Prophet or Darts.
- dbt experience for turning analysis logic into maintained models.
- Uplift modeling or heterogeneous treatment effect estimation.
Tech stack
Snowflake, dbt, Python 3.12, pandas, statsmodels, scikit-learn, PyMC, Jupyter and Hex, GrowthBook, Looker, GitHub. Replace this list with your real tools, including your experimentation platform if you have one.
What success looks like in 6 months
- You have run or reviewed experiments that led to at least two clear ship or kill decisions.
- Our experimentation process has a written standard for metrics, sample size and analysis that teams follow.
- At least one model or forecast you built is used in a recurring planning or operating meeting.
- Product and business leaders come to you before a decision, not only after it, to ask how to measure it.
How to apply and interview process
Send your resume or LinkedIn profile and a short note about an analysis that changed a decision. Our process has four steps: a 30-minute intro call, a technical conversation about past work, a practical case on experiment design or analysis, and a final conversation with the team and stakeholders you would work with. We aim to give feedback within a few days of each step.
Junior vs mid vs senior data scientist
Data science seniority shows in the questions a person chooses to answer and how much a decision maker can trust their conclusions without checking the work.
| Level | Scope | Typical experience | Key skills |
|---|
| Junior | Well-scoped analyses and dashboards with review from senior scientists | 0-2 years | SQL, Python or R, descriptive statistics, basic hypothesis tests, clear charts |
| Mid-level | Owns experiments and models for a product area, presents results to its team | 2-5 years | Experiment design, regression, tree-based models, metric definition, stakeholder communication |
| Senior | Sets methods and standards, frames strategic questions, advises leadership | 5+ years | Causal inference, variance reduction, Bayesian modeling, forecasting, influencing decisions |
Tips for writing a data scientist job description that attracts senior talent
- Name the decisions, not the algorithms. "Decide pricing changes across five markets" attracts senior data scientists. "Use machine learning to drive insights" attracts no one in particular.
- Separate data science from ML engineering. If the person must deploy models to production, say so and list the serving stack. If not, say that engineers own deployment.
- Describe your experimentation maturity. A team running hundreds of tests on a platform and a team about to run its first test need different people. Both are attractive to the right candidate.
- Say how close the role is to decision makers. Senior candidates want to know who reads their work and whether recommendations get acted on.
- Be honest about the data. If event tracking is patchy or definitions are disputed, mention it. Good data scientists can handle it, but they dislike surprises.
- Avoid the "unicorn" requirement list. Asking for deep learning, Spark, Tableau, causal inference and Kubernetes in one role signals an unclear job.
- Use a realistic case in the interview. A short experiment design or analysis problem from your own domain tests judgment better than algorithm trivia.
Skip the job post: hire a vetted senior data scientist
Finding a data scientist who combines solid statistics with clear business judgment can take months. Ryz Labs can match you with senior data scientists from Latin America who work on your team, work in your warehouse, notebooks and standups, and keep hours within ±1h of US time zones. Only the top 1% of the engineers and scientists we interview make it through our vetting, which covers statistics, experiment design, modeling and communication.
Our staff augmentation model lets you add one data scientist or a few. Ryz engineers join your team and report to your leads. Talk to us to scope your team. If the work grows into an AI system you need built and run in production, Ryz AI pod teams, dedicated pods of senior engineers that include a tech lead, ML and backend engineers, build it inside your cloud and repos alongside your team. Hiring on your own? Our data scientist interview questions cover what we test.
FAQ
What is the difference between a data scientist and a machine learning engineer?
A data scientist uses statistics and modeling to answer questions and guide decisions. A machine learning engineer builds the systems that train, serve and monitor models in production. Some roles mix both, but your job description should say which outcome matters most.
Does a data scientist need a PhD?
No. A PhD can help for research-heavy roles, but most product and business data science roles need strong statistics, SQL, Python or R and good judgment, which many people build through work experience. Use "or equivalent experience" in your requirements.
Should a data scientist job description list deep learning?
Only if the role involves it. Most decision-focused data science uses regression, tree-based models, time series and causal methods. Listing deep learning when it is not needed attracts candidates who want a different job.
Questions we didn't answer? Email info@ryzlabs.com.