CV and ATS guide for data scientists
They look for a decision you changed: a model in production, a metric that moved, or an analysis a business team used. “Built models in Python” without a use is invisible.
Common ATS gaps
- SQL / PostgreSQL on the JD while the CV only says “data wrangling.”
- Skills stuffed with every library from a bootcamp, none appearing in experience.
- Research-only wording when the JD wants production or stakeholder delivery.
Keywords that only count with proof
Python, SQL, and the warehouse or notebook stack you used — plus the decision or KPI. If the JD wants Gurobi or a named optimiser and you only used a generic solver, say the solver; do not rename it.
Common questions
- Do Kaggle ranks belong on a DS CV?
- Only if you can say what you did and what it proved. A rank without a method or dataset is weaker than one production analysis.
- Should I list every ML algorithm?
- No. Name the methods that appear in your bullets. ATS and humans both treat a long algorithm laundry list as noise.