Data Scientist Job Description

A copy-paste template for data science roles. Covers analytics, ML, and research tracks with 2025 salary benchmarks.

Data science hiring is notoriously difficult. The title means different things at different companies - from SQL analysts to ML researchers. Candidates are skeptical of vague postings because they've been burned before.

This template is designed to be specific: what will they actually work on, what tools will they use, and what impact will they have. Clarity attracts the right candidates and filters out the wrong ones.

The Template

Data Scientist

[Company Name] · [Location / Remote] · Full-time

About the Role

We're looking for a data scientist to join our [team name - e.g., Growth, Product Analytics, Risk, Pricing] team. You'll work on [specific problem area - e.g., user retention models, fraud detection, demand forecasting] that directly impacts [business outcome].

This isn't a role where you'll clean data in isolation. You'll work closely with [product managers / engineers / business stakeholders] to identify opportunities, design experiments, and deploy models that [specific impact - e.g., drive $X in revenue, reduce churn by Y%, improve Z by W%].

What You'll Do

  • Build and deploy [type of models - e.g., predictive models, recommendation systems, NLP pipelines] that [outcome]
  • Design and analyze A/B tests to measure impact of product changes
  • Partner with [engineers / product / business] to translate findings into actionable decisions
  • Develop dashboards and reporting to track key metrics for [team/domain]
  • Identify opportunities for automation and ML where it makes sense (not everywhere)
  • Present findings to stakeholders, including [leadership / executives / cross-functional teams]

What We're Looking For

Must-haves:

  • [X]+ years of experience in data science, analytics, or related quantitative field
  • Strong SQL skills - you can write complex queries without hand-holding
  • Proficiency in Python (pandas, scikit-learn) or R for analysis and modeling
  • Experience with statistical methods: hypothesis testing, regression, classification
  • Ability to communicate findings clearly to non-technical stakeholders
  • Track record of work that influenced real business decisions

Nice-to-haves (not required):

  • Experience with [specific tools relevant to role - e.g., Spark, Airflow, dbt, Looker]
  • Familiarity with deep learning frameworks (PyTorch, TensorFlow) for relevant use cases
  • Background in [domain - e.g., fintech, healthcare, marketplace, e-commerce]
  • Graduate degree in quantitative field (stats, math, physics, economics, CS)

Our Data Stack

[Be specific - data scientists want to know exactly what they'll work with.]

  • Warehouse: [e.g., Snowflake, BigQuery, Redshift, Databricks]
  • Orchestration: [e.g., Airflow, Dagster, Prefect]
  • Transformation: [e.g., dbt, Spark, custom Python]
  • ML Platform: [e.g., MLflow, SageMaker, Vertex AI, or "building it"]
  • BI/Viz: [e.g., Looker, Tableau, Mode, Hex, Metabase]
  • Experimentation: [e.g., internal platform, Statsig, Eppo, Optimizely]

The Team

You'll join [team name], a [size]-person team focused on [team mission]. The team includes [composition - e.g., 3 data scientists, 2 data engineers, 1 analytics engineer]. You'll report to [title - e.g., Head of Data Science, Director of Analytics].

[Optional: How the DS team works - embedded in product teams vs centralized, research vs applied focus, collaboration patterns.]

Compensation & Benefits

  • Salary: $[X] - $[Y] depending on experience
  • Equity: [Yes/No] - [brief description if yes]
  • [Benefit 1 - e.g., Health/dental/vision insurance]
  • [Benefit 2 - e.g., Flexible PTO / Unlimited PTO]
  • [Benefit 3 - e.g., 401k matching / Retirement plan]
  • [Benefit 4 - e.g., Learning budget / Conference attendance]

About [Company Name]

[2-3 sentences about what the company does and why data matters here. Focus on interesting problems, scale of data, or business impact - not generic mission statements.]

Interview Process

  1. Application review (we respond within [X] days)
  2. 30-minute intro call with recruiter or hiring manager
  3. Technical screen: [format - e.g., SQL exercise, take-home analysis, live coding]
  4. Virtual onsite (4-5 hours):
    • Case study / problem-solving session
    • Technical deep dive on past projects
    • Cross-functional interview (working with stakeholders)
    • Culture / values conversation
  5. References + offer

[Company Name] is an equal opportunity employer. We value diversity and don't discriminate based on race, religion, color, national origin, gender, sexual orientation, age, marital status, veteran status, or disability status.

Level Variations

Adjust the template based on seniority. Here's what changes at each level:

Junior / Entry-Level

0-2 years experience

  • Remove or soften years of experience requirement
  • Focus on fundamentals: SQL, Python/R basics, statistics
  • Emphasize mentorship and structured learning
  • Expect more guidance on project scoping
  • New grads, bootcamp grads, career changers welcome

Change: "Experience with statistical analysis" instead of "5+ years of data science experience"

Senior

5-8+ years experience

  • Define the analytical agenda for a domain
  • Mentor junior data scientists
  • Influence product and business strategy
  • Design systems and frameworks, not just analyses
  • Navigate ambiguous problems with minimal guidance

Add: "Define the data science roadmap for [domain]" and "Mentor junior team members"

Staff / Principal

8+ years experience

  • Set technical direction for DS org
  • Org-wide impact across multiple teams
  • Build DS capabilities and best practices
  • Executive communication and influence
  • May include management responsibilities

Add: "Shape the data science strategy across the company" and "Represent DS in executive forums"

Data Scientist vs ML Engineer

The biggest source of confusion in DS hiring. Be clear about which role you're actually hiring for:

Dimension Data Scientist ML Engineer
Primary output Insights, recommendations, models Production systems and pipelines
Time horizon Analysis to decision (days/weeks) System reliability (months/years)
Key skills Statistics, experimentation, communication Engineering, MLOps, system design
Collaborates with Product, business, leadership Engineering, infrastructure, DS
Success metric Business impact of decisions Model performance in production
Background Stats, econ, physics, social science CS, software engineering

Hiring a Data Scientist

  • Emphasize business context and stakeholder work
  • Focus on statistical rigor and experimental design
  • Look for communication skills in interviews
  • Case studies should involve ambiguous problems
  • Production engineering is nice-to-have, not must-have

Hiring an ML Engineer

  • Emphasize system design and scale
  • Focus on engineering best practices
  • Look for production experience in interviews
  • System design interviews are essential
  • Business communication is nice-to-have

Hybrid roles exist - especially at smaller companies where one person does both. If that's you, be explicit: "This is a hybrid DS/MLE role" and list responsibilities for each.

2025 Salary Benchmarks

US market rates for data scientists. Adjust for location, company stage, and specialization (ML-heavy roles trend higher).

Level Startup (Seed-A) Growth Stage Big Tech / FAANG
Junior (0-2 yr) $90K - $120K $110K - $150K $150K - $200K
Mid (2-5 yr) $120K - $170K $150K - $200K $200K - $300K
Senior (5+ yr) $160K - $220K $190K - $280K $280K - $420K
Staff (8+ yr) $200K - $300K $250K - $380K $380K - $550K

Note: Big Tech total comp includes significant RSUs (often 40-60% of package). Startups may offer larger equity grants with highly variable outcomes. Specialized roles (NLP, CV, ML research) can command 10-20% premiums. These figures represent total comp, not base salary alone.

Writing Tips for Data Science JDs

01

Define "data science" for your company

The title is meaningless without context. Are you hiring for analytics, ML engineering, research, or something hybrid? State it upfront.

02

Show the data stack

Data scientists care about tools. Snowflake vs Redshift matters. Having MLflow vs "we'll figure it out" matters. Be specific.

03

Describe the data, not just the role

What data do you have? How much? How clean? "10M daily events" or "messy legacy warehouse" sets expectations accurately.

04

Show real impact examples

"Built model that increased conversion 15%" beats "apply ML to business problems." Past wins signal future opportunity.

05

Be honest about the data maturity

Early-stage data infra? Say so. Some candidates want to build from scratch. Others want established pipelines. Don't mislead.

06

Don't require PhDs unless necessary

Most DS work doesn't require a PhD. If you do need research-level depth, explain why. Otherwise, you're filtering out great candidates.

Common Mistakes to Avoid

Avoid

"Looking for a unicorn who can do ML, analytics, data engineering, and manage a team"

Better

"Looking for a data scientist focused on [specific area] who will collaborate with our data engineering team"

Unicorn job descriptions attract no one. Be realistic about what one person can do well.

Avoid

"Must have PhD in Computer Science, Statistics, or related field"

Better

"Strong quantitative background - relevant experience matters more than credentials"

Unless you're hiring for a research role, PhDs aren't necessary. Many excellent data scientists are self-taught or have non-traditional backgrounds.

Avoid

"Experience with TensorFlow, PyTorch, Keras, scikit-learn, XGBoost, LightGBM, CatBoost, Spark MLlib, H2O..."

Better

"Experience with ML frameworks (we use PyTorch and scikit-learn, but similar experience transfers)"

Laundry lists signal that you don't know what you need. Good data scientists learn new tools quickly.

Avoid

"Use cutting-edge AI to revolutionize our industry"

Better

"Build demand forecasting models to reduce inventory costs by 20%"

Vague AI hype attracts hype-driven candidates. Specific problems attract problem-solvers.

Real-World Examples

Companies known for clear, effective data science job descriptions:

Airbnb

Known for: Clear distinction between DS tracks (analytics, algorithms, inference), strong impact focus

Spotify

Known for: Specific team context, research vs applied clarity, compelling problem descriptions

Stripe

Known for: Business impact focus, clear scope, honest about challenges and data maturity

DoorDash

Known for: Detailed technical requirements, clear ML vs analytics distinction, scale context

Study what works, but write a JD that sounds like your company, not a copy of theirs.