“Data scientist” is one job title sitting on top of a dozen different skill stacks. If you can write Python, run a regression and explain a messy result to someone who doesn’t care about p-values, you qualify for far more than the title on your degree. This post covers the jobs a data scientist can do, how each one differs from the original, and roughly what it pays.
Salary figures are US ranges pulled from Bureau of Labor Statistics data and public job listings. The BLS puts the median for data scientists at about $112,000. Treat every range below as a ballpark: city, industry and company size move these numbers by tens of thousands.
Contents
- The 12 roles at a glance
- Adjacent technical roles
- Analytical roles
- Non-obvious pivots
- Beyond the corporate job
- How AI tools are changing these jobs
- How to choose your path
- FAQ
The 12 roles at a glance
| Role | Group | Typical US pay | Realistic for juniors? |
|---|---|---|---|
| Machine Learning Engineer | Technical | $130k-$200k+ | Sometimes |
| Data Engineer | Technical | $115k-$175k | Yes |
| Analytics Engineer | Technical | $100k-$155k | Yes |
| Data Analyst | Analytical | $65k-$100k | Yes |
| BI Developer | Analytical | $85k-$125k | Yes |
| Product Analyst | Analytical | $95k-$150k | Yes |
| Statistician / Quant | Analytical | $100k-$300k+ | Statistician yes, quant rarely |
| AI Research Scientist | Pivot | $150k-$300k+ | Rarely (usually needs a PhD) |
| NLP Engineer | Pivot | $125k-$190k | Sometimes |
| Data Science Manager | Pivot | $160k-$250k | No |
| Data Science Consultant | Pivot | $90k-$180k | Sometimes |
| Data Journalist | Pivot | $55k-$95k | Yes |
Adjacent technical roles
These are the closest relatives. Same toolbox, different center of gravity: less exploring, more building things that have to keep running.
1. Machine Learning Engineer
What it is: You take a model that works in a notebook and make it work in production: serving it, monitoring it, retraining it when the data drifts.
How it differs from a data scientist: A data scientist asks whether a model should exist. An ML engineer asks how it runs at 10,000 requests a second without falling over. The code is closer to software engineering than analysis.
Core skills: Python, PyTorch or TensorFlow, Docker, cloud platforms, MLOps tooling, solid software engineering habits like testing and code review.
Pay: $130k-$200k+, higher at big tech and AI labs.
Suits: People who like their code to ship and don’t mind that “the model is 2% better” matters less than “the pipeline hasn’t broken.”
2. Data Engineer
What it is: You build the pipelines that move data from apps, sensors and databases into a warehouse where anyone can use it. Without this role, data scientists spend half their week cleaning CSVs.
How it differs from a data scientist: Almost no modeling. You care about SQL performance, schema design and whether the overnight job finished.
Core skills: SQL, Python, Spark, Airflow or similar orchestrators, cloud warehouses like Snowflake or BigQuery.
Pay: $115k-$175k.
Suits: Data scientists who discovered they enjoy the plumbing more than the analysis. It’s also one of the easiest lateral moves, since you already know what downstream users need.
3. Analytics Engineer
What it is: A newer role that sits between data engineering and analysis. You write the tested, documented SQL models (usually in dbt) that turn raw tables into clean datasets the whole company trusts.
How it differs from a data scientist: You produce reliable definitions of “active user” or “revenue,” not predictions. Think software practices applied to analytics.
Core skills: Advanced SQL, dbt, version control, data modeling, enough business sense to argue about what a metric should mean.
Pay: $100k-$155k.
Suits: Detail-oriented people who are tired of three dashboards showing three different revenue numbers.
Analytical roles
If you liked the questions more than the models, this group keeps you close to the business with less machine learning.
4. Data Analyst
What it is: You answer specific business questions with data: why did sign-ups drop in March, which stores underperform, what did the pricing test actually show.
How it differs from a data scientist: More descriptive, less predictive. The work leans on SQL, spreadsheets and visualization rather than modeling pipelines, and the turnaround is days rather than weeks.
Core skills: SQL, Excel, a BI tool, basic statistics, clear writing.
Pay: $65k-$100k. This is the lowest-paying role on the list, which is why it works as a fallback or a first rung rather than a destination for most experienced data scientists.
Suits: Juniors and career changers. It’s the most available entry point in the field.
5. BI Developer
What it is: You build and maintain the dashboards and reporting systems executives open every Monday morning.
How it differs from a data scientist: You design for repeated use by non-technical people. A model gets used once; a dashboard gets opened 400 times.
Core skills: Tableau, Power BI or Looker, SQL, data warehousing, an eye for what a manager needs to see in five seconds.
Pay: $85k-$125k.
Suits: People who enjoy design and want stable, well-defined work.
6. Product Analyst
What it is: You embed with a product team at a software company and work out how people use the product: funnels, retention, A/B tests, feature adoption.
How it differs from a data scientist: Experimentation is the center of the job. Many companies use “product data scientist” and “product analyst” for overlapping work, and the main difference is often the title.
Core skills: SQL, experiment design and causal inference basics, Python or R, product intuition.
Pay: $95k-$150k.
Suits: Data scientists who like fast feedback and working next to designers and engineers.
7. Statistician or Quantitative Analyst
What it is: Two jobs that share a foundation. Statisticians design studies and analyze data in pharma, government, agriculture and public health. Quants build pricing and trading models at banks and hedge funds.
How it differs from a data scientist: More mathematical rigor, less engineering. A clinical trial statistician lives by pre-registered analysis plans; a quant lives by stochastic calculus and low-latency code.
Core skills: Probability, inference, experimental design, R or SAS for statisticians; C++, Python and time-series math for quants.
Pay: The BLS median for statisticians is about $104,000. Quant pay is a different universe, from $150k to well over $300k with bonuses.
Suits: People with a strong math background who prefer being right to being fast.
Non-obvious pivots
These take extra specialization or a different kind of career, but the data science foundation carries over directly.
8. AI Research Scientist
What it is: You develop new model architectures and training methods, then publish or build them into products. Research labs and the research arms of big tech companies hire most of these people.
How it differs from a data scientist: You’re inventing methods, not applying existing ones. The BLS category for computer and information research scientists, which includes AI researchers, reports a median above $140,000.
Core skills: Deep learning theory, strong linear algebra and optimization, a publication record, usually a PhD.
Pay: $150k-$300k+.
Suits: Academics at heart. Not a realistic first job for most people.
9. NLP Engineer
What it is: You build systems that work with human language: search, chatbots, document classification, summarization, translation. Since large language models took over, much of the work is adapting, evaluating and deploying them.
How it differs from a data scientist: A narrow specialty with its own tooling. You’ll spend more time on tokenization, retrieval and evaluation than on general-purpose tabular modeling.
Core skills: Python, transformers libraries, vector databases, prompt and evaluation design, linguistics helps.
Pay: $125k-$190k.
Suits: Data scientists who’ve already done text projects and want to specialize in the hottest part of the field.
10. Data Science Manager
What it is: You lead a team of analysts and data scientists: set priorities, hire, review work and translate between the team and executives.
How it differs from a data scientist: You write less code every year. Your output is other people’s output, plus the stakeholder conversations that decide which projects get funded.
Core skills: Everything from the technical track, plus coaching, prioritization and the ability to say no.
Pay: $160k-$250k.
Suits: People with five or more years of experience who find they enjoy unblocking colleagues more than building models themselves.
11. Data Science Consultant
What it is: You solve data problems for multiple clients, either inside a firm like Deloitte or as an independent freelancer.
How it differs from a data scientist: Every project is different, deadlines are short, and half the job is working out what the client actually needs. The data is rarely as clean as promised.
Core skills: Broad technical range, client communication, scoping, and the ability to produce something usable in weeks.
Pay: $90k-$180k as an employee; experienced freelancers charge $100-$250 an hour.
Suits: Generalists who get bored easily.
12. Data Journalist
What it is: You find stories in datasets and tell them with charts, maps and code. Newsrooms like The New York Times, The Guardian and ProPublica employ people who scrape public records, analyze them and build interactive pieces.
How it differs from a data scientist: The audience is the public, and the product is a story. Methods stay simple on purpose, because readers need to understand them.
Core skills: Python or R, SQL, data visualization, web scraping, and enough writing ability to carry a narrative.
Pay: $55k-$95k. Newsroom pay is the lowest on this list.
Suits: Data scientists who want their work to matter outside a company’s quarterly numbers.
Beyond the corporate job
Most lists assume a data scientist wants a tech-company badge. Several of the roles above exist elsewhere.
Healthcare and biotech. Hospitals, insurers and pharma companies hire statisticians, analysts and ML engineers for clinical trials, imaging models and claims analysis. The regulatory overhead is heavy, and domain knowledge counts for as much as technical skill.
Finance. Quants, risk analysts and fraud-detection engineers are the obvious paths. Fraud teams in particular hire for ML engineering skills and pay well.
Sports. Professional teams now run analytics departments. Pay is lower than in finance, and the competition for seats is intense.
Climate and environment. Government agencies, research institutes and energy companies need people to model weather, emissions and grid demand. The pay is usually lower than in tech.
Academia and government. A research scientist position at a university or national lab pays less but offers independence. Government agencies hire statisticians and operations research analysts with stable contracts and clear pay scales.
Freelancing. Consulting and contract work is open to anyone with a portfolio and a few clients. Income swings are the cost of the flexibility.
How AI tools are changing these jobs
Code assistants now write the first draft of most SQL queries and plotting code. That shifts value in two directions. Routine reporting and basic analysis get cheaper, which pressures entry-level analyst work. Meanwhile, judgment about what to measure, whether a result is trustworthy and how to run a clean experiment becomes the scarcer skill.
Roles tied to building and operating AI systems (ML engineer, NLP engineer, data engineer) are growing the fastest. Pure reporting roles are the most exposed. If you’re early in your career, learn to use these tools quickly, but keep practicing the fundamentals so you can tell when the output is wrong.
How to choose your path
Start with what you’re strongest at, not what pays most.
- Strong coding, weaker statistics: ML engineer, data engineer or analytics engineer.
- Strong statistics, weaker engineering: statistician, quant, product analyst or AI research.
- Strong communication: data analyst, BI developer, consultant, manager or data journalist.
- Want the highest ceiling: ML engineer, AI researcher, quant or manager. All of them demand either deep specialization or years of experience.
- Want the easiest way in: data analyst, analytics engineer, BI developer or data engineer. These have the most openings at junior level.
The skill overlap is large. SQL and Python appear in nearly every role above, and moving between adjacent roles (analyst to analytics engineer, data scientist to ML engineer) usually takes six to twelve months of focused project work rather than a new degree. The jobs a data scientist can do are less a ladder than a map, and most people move sideways at least once.
FAQ
What is the highest paying job for a data scientist? Quant roles and AI research scientist positions top the list, often exceeding $300k with bonuses or equity. Data science managers and senior ML engineers at large tech companies follow.
Is machine learning engineer better than data scientist? Neither is better. ML engineers earn slightly more on average and work closer to software engineering. Data scientists spend more time on analysis and experimentation.
Can a data scientist become a data analyst? Yes, and it’s an easy move. You’ll likely take a pay cut, but the work-life balance is often better and the role is easier to find.
Which role is easiest to get without experience? Data analyst, followed by BI developer and data engineer. Analytics engineering is also open to juniors who know SQL well.
Do I need a PhD for any of these? Only for AI research scientist, and for some quant and pharma statistician roles. The other ten are open to bachelor’s and master’s graduates.
