Data Scientist Job Description
A complete data scientist job description template — what the title meant in 2015 versus now, plus the ATS keywords real postings screen for.
"Data Scientist" has drifted further from its original meaning than any other title on this list. In 2015 it described a fairly specific profile — usually a statistics, math or computer science background, writing bespoke models to answer questions no off-the-shelf tool could touch. Managed machine learning platforms and general-purpose model APIs have since absorbed a lot of that work, and the title expanded to cover everyone left standing near a dataset.
The result is that two "Data Scientist" postings can describe almost unrelated jobs. One is a data analyst with Python: SQL, dashboards, the occasional regression, and a job title that sounds more technical than the daily work is. The other is close to applied research: designing experiments, building and validating models, and reasoning carefully about when a result is real versus noise. The posting's language is a weak signal for which one you're looking at — the interview loop is a better one. A take-home built around a messy dataset and a stakeholder memo points one way; a system-design or statistics-fundamentals interview points the other.
Two neighboring titles are worth knowing to place yourself against. An analytics engineer builds and tests the tables a data scientist should be able to trust without re-deriving them from raw events — if a posting expects you to also own that transformation layer, the team likely doesn't have one yet. A machine learning engineer takes a model past the prototype and into a production system with monitoring and retraining — at most mid-sized companies, a data scientist partners with that role rather than owning it. If you're genuinely torn between the two titles, ask whether you're more drawn to the question or to the system that keeps answering it after you've moved on to the next one.
Bellamyre · Remote (US)
Full-time · Remote
$115,000 – $155,000
About the role
Bellamyre is hiring a Data Scientist to join the modeling team supporting our subscription and usage-based products. You'll design and run experiments, build models that predict churn and usage patterns, and work closely with the analytics engineering and machine learning engineering teams who prepare the data you use and productionize what you build.
This is a fully remote, US-based role. The team runs async-first with a few standing meetings a week, and you'll be expected to write findings up clearly enough that someone in a different time zone can act on them without a live walkthrough.
What you'll do
- Design and run experiments — A/B tests and quasi-experiments — to answer specific product and pricing questions.
- Build and validate statistical and machine learning models, from a simple baseline through to something worth shipping.
- Partner with analytics engineering to source clean, tested data rather than re-deriving it from raw events yourself.
- Hand modeling work bound for production to machine learning engineering with enough documentation that they can own it.
- Communicate findings — including negative or inconclusive ones — in writing a non-technical stakeholder can act on.
- Monitor the accuracy of models already in production, flagging drift before it becomes someone else's incident.
- Push back on requests for a model when a simpler analysis would answer the question just as well.
- Contribute to the team's shared library of reusable analysis code rather than rewriting the same query from scratch.
What we're looking for
- Three or more years applying statistics or machine learning to real business problems, not just coursework.
- Strong Python, including pandas and at least one modeling library such as scikit-learn.
- SQL fluent enough to pull and shape your own data without waiting on someone else.
- A solid grounding in experiment design and statistical inference — you can explain why a result might be noise.
- Experience presenting analysis to non-technical stakeholders and adjusting the message, not just the slides, for the audience.
- Comfort working with imperfect, real-world data rather than a cleaned textbook dataset.
Nice to have
- Experience with causal inference methods beyond a standard A/B test, such as difference-in-differences or matching.
- Familiarity with cloud data platforms such as Snowflake or BigQuery.
- Exposure to shipping a model into production, even if you weren't the one who owned the deployment.
- A graduate degree in statistics, economics, computer science or a related quantitative field.
- Experience with subscription, usage-based or marketplace business models specifically.
Benefits
- Medical, dental and vision coverage, with employee premiums covered in full.
- 401(k) with a 4% company match, vested immediately.
- Fully remote within the US, with a yearly team offsite.
- $2,500 annual learning budget, usable on conferences, courses or books.
- Twenty-one days of paid time off plus company holidays.
Salary range
As posted for this sample role. Real pay varies by employer, location and experience.
$115,000–$155,000/ yr
ATS keywords for this role
The applicant tracking system (ATS) — the recruiting software a hiring team searches and filters applicants with — will screen for these. Weight shows how central each one is to this specific posting.
Required and central (4)
Important (8)
Mentioned in passing (5)
Frequently asked questions
Is a PhD required to become a data scientist?
No, and it's required less often than it used to be. A master's degree or strong self-directed experience — competitions, a portfolio of real analyses — is enough at most companies; a PhD matters more for research-heavy roles at large tech companies or in specialized domains like healthcare or genomics.
What's the difference between a data scientist and a machine learning engineer?
Roughly, the question versus the system. A data scientist is usually measured on whether an analysis or model answers the business question correctly; a machine learning engineer is measured on whether that model keeps running correctly in production, at scale, months later. Plenty of people move between the two over a career.
Why do data scientist postings vary so much in what they ask for?
Because the title absorbed several older, more specific roles as companies of every size adopted it. A ten-person startup's "Data Scientist" may really need a data analyst who can code; a large company's may be one of several specialists on a modeling team. Read the requirements and the interview process, not the title, to know which one you're looking at.
This was a sample. Your resume should be tailored to the real thing.
Rezi Ninja reads an actual job posting and rewrites your resume to match it, with a Ninja Score so you know it lands before you hit send. Free to start, with the AI usage included.