Built machine learning models to help the team make better business decisions.
✓ AFTER HAYVEN
Built a churn prediction model (XGBoost, 91% AUC) that identified at-risk accounts 30 days earlier — enabling targeted outreach that reduced churn by 18% and saved $2.3M ARR annually.
AUC, F1, precision/recall, RMSE — always include model performance metrics. "Built a model" means nothing. "Built a model with 91% AUC" means everything.
Business impact over technical complexity
Hiring managers care more about what the model did for the business than how it worked. Lead with the outcome, then mention the method.
Stack and tools matter a lot
Python, SQL, R, TensorFlow, PyTorch, Spark, Databricks, Airflow, dbt, Snowflake — ATS systems for DS roles scan for these. List them in a dedicated Skills section.
Scale gives context
How large was the dataset? How many predictions per day? How much compute? Scale tells recruiters whether you can handle their data environment.
Deployed vs. prototype
Distinguish between models you deployed to production and experiments that stayed in notebooks. Production experience is significantly more valuable.
Domain expertise compounds
Healthcare, fintech, e-commerce, NLP, computer vision — domain-specific DS experience is a major differentiator. Name your domain explicitly in your summary.
Get your DS resume reviewed in 60 seconds
ATS score, bullet rewrites, and a prioritized action plan — free.