Internship Data Scientist resume example

• Under supervision, separated training and evaluation data by time and customer, inspected leakage in derived features, and compared model performance across cohorts before recommending an experiment.
• The team defined a cancellation-risk cohort and trained a calibrated gradient-boosted model; I supported research preparation, validation, and documentation under mentor review. The team identified the highest-risk account group where churn was concentrated; confirmed with my mentor that my contribution was limited to the assigned support and validation work.
• The team designed a holdout test for a targeted retention offer across 48,000 accounts; I supported research preparation, validation, and documentation under mentor review. The team created attributable pipeline from the targeted work; confirmed with my mentor that my contribution was limited to the assigned support and validation work.
• Under supervision, built a Python machine learning baseline for repeat-purchase prediction, split observations by customer and time, and compared calibration and recall with a simple statistical baseline before presenting the model recommendation.
• Under supervision, prepared an experiment readout with confidence intervals, cohort-level diagnostics and a data visualization of the pre-test trend; checked sample-ratio mismatch and explained when the evidence did not justify rollout.
• Under supervision, owned a forecasting evaluation using time-based splits in Python, reducing holdout error from 19% to 15% against the same seasonal baseline and recording cohort failures.
• Under supervision, published a Tableau evaluation dashboard using SQL-validated inputs, resolving differences between data-pipeline totals and the statistics presented to decision makers.

• In a mentor-reviewed simulation, used SQL and Python to study seller retention, separating new-seller cohorts from established shops and checking seasonality; reported uncertainty rather than treating a correlation as a causal effect.
• As a second supervised exercise, built Looker dashboards from the research dataset with documented cohort definitions and drill-through samples; reconciled dashboard totals to the source queries before sharing recommendations.
• Under supervision, the team reconciled three conflicting definitions of active supply into one governed metric; I supported research preparation, validation, and documentation under mentor review
• Generated synthetic source records with duplicates, late arrivals and corrected values; wrote assertions for row counts and key uniqueness, and recorded the expected effect of each case on the reported metric.
• Compared the analytical output with a manually calculated reference table, traced differences to a transformation step, and retained a data dictionary and rerun instructions alongside the corrected query.
• Owned the synthetic-data validation using a manually calculated reference, resolved duplicate-key inflation and completed a notebook that reproduces the corrected totals.

Relevant coursework: Algorithms, operating systems, databases, computer networks

• Published an independent methods note using reproducible queries, explaining the data grain, excluded records and sensitivity of the result to a changed denominator.