Should a data scientist CV include Kaggle projects?

Only if they show a skill your work does not, or if you have little work experience. Describe the problem and result, not just the ranking.

How do I show the impact of a model?

Compare it with the previous approach, say who used it and describe the effect you actually observed.

Do I need deep learning on my data scientist CV?

Only if the posting asks for it. Many data scientist roles focus on statistics, experiments and classical models.

Data

Data Scientist CV: what gets you shortlisted

Data scientist postings vary widely, from experimentation to machine learning in production. Screeners want to see which kind you are, what models or analyses you built, how you judged them and whether they made it into real use.

Get my free score

One-time analysis • $1.99 • No subscription

What employers look for in a Data Scientist CV

Screeners check these first. Each one should be visible in your job entries, not only in a skills list.

Python with the usual libraries, and SQL for getting your own data.
Models or statistical analyses that reached real use, with how they were evaluated.
Experimentation skills: designing tests, choosing metrics and reading results correctly.
Communicating findings to people who are not data scientists.
For ML-heavy roles, experience getting models into production with engineers.

Skills Data Scientist postings often ask for

Listing a skill is not the same as showing it. Resymetry counts a skill as matched only where your CV shows you used it.

Python
SQL
pandas
scikit-learn
Statistics
Experiment design
Machine learning
Jupyter
Data visualisation
PyTorch or TensorFlow

4 common Data Scientist CV mistakes and how to fix them

Listing algorithms instead of problems.
Fix: Start with the business problem, then the approach and the result.
Reporting accuracy with no baseline.
Fix: Compare against what existed before, and give the business effect where you know it.
Course and competition projects outweigh work experience.
Fix: Lead with work. Keep only projects that show something your jobs do not.
Not saying whether a model was ever used.
Fix: State whether it went to production, informed a decision or stayed a prototype.

Weak and strong CV bullets

Illustrative examples, not real candidates. Use your own work and only numbers you can back up.

Weak: Built machine learning models.
Stronger: Built a gradient-boosted churn model that beat the rule-based baseline by 12 points of recall; the retention team now uses its weekly list.
Weak: Ran experiments.
Stronger: Designed the experiment for a new recommendation widget, choosing guardrail metrics that caught a drop in repeat purchases before full rollout.
Weak: Presented insights to stakeholders.
Stronger: Presented demand-forecast results to the operations leads, who used them to set staffing for the holiday peak.

Know your chances for Data Scientist roles

Your score and your biggest problems are free. The full report is a one-time $1.99.