Data Scientist resume example
By Bhagirath Sheela, founder of JobCatch ·
Data science resumes often drown in model names. Hiring managers want to know which business problems you solved with data, how you measured the result, and whether your models ran in production — or only in a notebook.
What recruiters look for
- Problems solved: forecasting, churn, fraud, recommendations, pricing
- Results measured against a baseline, in money or accuracy
- Python, SQL and the ML libraries, plus statistics depth
- Deployment: models that are used, monitored and retrained
Section order
With experience
- Summary
- Skills
- Experience
- Projects
- Education
- Certifications
As a fresher
- Summary
- Education
- Skills
- Projects
- Internships
- Certifications
Summary examples
Fresher
Data scientist with an M.Sc. in statistics, an internship building a churn model for a telecom company, and two Kaggle competitions finished in the top 10%.
Experienced
Data scientist with four years in lending and payments. Built the fraud model that blocks ₹3 crore of fraudulent transactions a month at a 0.4% false-positive rate.
Skills, spelled the way postings spell them
An applicant tracking system matches the words in the job posting, and it does not reliably treat two spellings as one skill. The left column is the form most postings use; the right, other spellings people type. Use the posting's own spelling for each application, and list only what you can discuss in an interview.
| Skill | Also written as |
|---|---|
| Python | Python 3 |
| SQL | Structured Query Language |
| Machine learning | ML |
| Statistics | Statistical analysis |
| Pandas | — |
| NumPy | — |
| Scikit-learn | sklearn, Scikit learn |
| Deep learning | — |
| PyTorch | Py Torch |
| Natural language processing | NLP |
| Apache Spark | Spark, PySpark |
| Data analysis | Data analytics |
Also worth listing if you have them: A/B testing, Feature engineering.
Bullet examples
Each bullet says what you did and what changed because of it. Change the numbers to your own; a recruiter will ask about every one.
- Built a gradient-boosted fraud model that blocks ₹3 crore of fraud a month at a 0.4% false-positive rate.
- Replaced rule-based credit limits with a model that cut 90-day defaults by 11% at the same approval rate.
- Designed the A/B framework used to test every pricing change, with power calculations built in.
- Worked with ML engineers to move three models from notebooks to a monitored scoring service.
- Built an NLP classifier that routes 70% of support tickets automatically.
For freshers: projects and internships
- Internship: built a churn model for a telecom company; the top decile captured 38% of churners.
- Finished in the top 10% of two Kaggle competitions on tabular data.
- M.Sc. dissertation on forecasting monsoon rainfall with gradient boosting and ARIMA.
Mistakes to avoid
- Accuracy with no baseline or business number. "92% accuracy" means little; "cut defaults 11%" means a lot.
- Listing every algorithm you have read about. List the ones you used and can defend.
- Only notebooks. Show at least one model that someone used, or describe how yours would be deployed.
What an ATS reads from this sample
An applicant tracking system turns the PDF into plain text before anyone reads it. This is that text for the sample above, extracted by the same code as "What the ATS reads" in the app. Everything is in reading order, under the right heading, with the name and email intact — which is what a single-column template is for.
Siddharth Rao Data Scientist [email protected] • +91 98765 43210 • Bengaluru, Karnataka • linkedin.com/in/siddharthrao Summary Data scientist with four years in lending and payments. Built the fraud model that blocks ₹3 crore of fraudulent transactions a month at a 0.4% false-positive rate. Skills & Tools Python SQL Machine learning Statistics Pandas NumPy Scikit-learn Deep learning PyTorch Natural language processing Apache Spark Data analysis A/B testing Feature engineering Experience Data Scientist, Lendmark Finance Mar 2022 – Present Bengaluru Built a gradient-boosted fraud model that blocks ₹3 crore of fraud a month at a 0.4% false-positive rate. Replaced rule-based credit limits with a model that cut 90-day defaults by 11% at the same approval rate. Designed the A/B framework used to test every pricing change, with power calculations built in. Associate Data Scientist, Datapulse Analytics Jul 2020 – Feb 2022 Bengaluru Worked with ML engineers to move three models from notebooks to a monitored scoring service. Built an NLP classifier that routes 70% of support tickets automatically. Projects Ticket router NLP classifier that routes 70% of support tickets to the right team automatically. Education M.Stat., Statistics 2018 – 2020 Indian Statistical Institute · First class Languages English · Hindi
Interview questions to prepare
The app's interview prep for this sample and a plain Data Scientist posting. It works from the resume and the posting only, and tells you how to prepare, never what to say.
Questions every interview has
- Tell me about yourself. Two minutes: what you do now (Data Scientist at Lendmark Finance), one result you are proud of, and why this role is next.
- Why do you want this job? Read their website and one recent piece of news before the call. Name something specific.
- Why are you leaving your current job? Say what you are moving towards. Do not criticise your employer.
- What is your notice period, and what salary are you expecting? Decide your number before the interview, so you are not deciding it on the call.
About the skills the job asks for
- Tell me about a time you used Python. Pick one real example: what the problem was, what you did, and what changed.
- Tell me about a time you used SQL. Pick one real example: what the problem was, what you did, and what changed.
- Tell me about a time you used Pandas. Pick one real example: what the problem was, what you did, and what changed.
- Tell me about a time you used NumPy. Pick one real example: what the problem was, what you did, and what changed.
- Tell me about a time you used Machine learning. Pick one real example: what the problem was, what you did, and what changed.
- Tell me about a time you used Statistics. Pick one real example: what the problem was, what you did, and what changed.
About what the resume claims
- You wrote: "Built a gradient-boosted fraud model that blocks ₹3 crore of fraud a month at a 0.4% false-positive rate." Expect three follow-ups: how exactly, what was your own part, and how you know it worked.
- You wrote: "Replaced rule-based credit limits with a model that cut 90-day defaults by 11% at the same approval rate." Expect three follow-ups: how exactly, what was your own part, and how you know it worked.
- You wrote: "Designed the A/B framework used to test every pricing change, with power calculations built in." Expect three follow-ups: how exactly, what was your own part, and how you know it worked.
Questions to ask them
- What would a good first three months in this role look like?
- Who would I work with most closely?
- What are the next steps, and when should I expect to hear back?
Templates that suit this role
All of these are single-column and read cleanly by an applicant tracking system. Classic is free, with no watermark.
Questions
Do I need a master's degree to be a data scientist in India?
It is common but not required. Strong projects, statistics fundamentals and SQL can get you in, often through an analyst role first.
Should a data scientist resume include Kaggle?
Good ranks, yes, especially for freshers. Kaggle shows modelling skill; pair it with one project on a real business problem, which shows judgment.
Next: the Data Scientist cover letter the app drafts from this resume, and how to show Python, SQL, Statistics on a resume.
More data & ai resume examples: Data Analyst · Business Analyst · Data Engineer · Machine Learning Engineer · AI Engineer · Power BI Developer · SQL Developer