Rahul Maldonado

Summary

Statistics graduate with practical experience building predictive models and running experiments on real product data. Comfortable across the full workflow from SQL data pulls through model evaluation in scikit-learn. Looking to grow as a data scientist on a team that values rigorous experimentation and clear communication of results.

Skills

• Languages & Tools: Python (pandas, NumPy, scikit-learn), SQL, Jupyter, Git
• Statistics & ML: hypothesis testing, A/B test design, regression and classification models, feature engineering
• Core Competencies: data storytelling, exploratory data analysis, cross-functional collaboration

Experience

Junior Data Scientist — Cobalt Fitness (Aug 2025 – Present)

• Built a logistic regression churn model on 90,000+ member records, achieving 0.81 AUC and flagging at-risk members for a retention campaign
• Designed and analyzed an A/B test on the app’s onboarding flow, contributing to a 9% lift in 30-day retention
• Wrote SQL pipelines to pull and clean membership and usage data for weekly reporting, reducing manual pull time by 4 hours per week
• Presented model results and experiment findings to the product and marketing teams in biweekly reviews

Data Science Intern — Wavecrest Telecom (May 2024 – Aug 2024)

• Assisted in building a customer segmentation model using k-means clustering on usage and billing data for 50,000 customers
• Contributed feature engineering work that improved a churn prediction model’s precision by 6 percentage points
• Automated a recurring data quality check in Python, catching missing-value issues before they reached the modeling pipeline

Education

• B.S. in Statistics, Cedar Valley University, 2025
• Relevant coursework: Applied Machine Learning, Experimental Design, Bayesian Statistics, Databases

Certifications / Projects

• Kaggle Competition: Finished in the top 15% of a customer churn prediction competition using a gradient-boosted tree ensemble
• Capstone Project: Built an end-to-end pipeline predicting subscription cancellations from behavioral data, deployed as a lightweight Flask API for demo purposes