CareforWe – Data Science Intern

Role Overview

You will lead research into CKD risk factors and help design a predictive algorithm to estimate CKD risk levels. This involves combining clinical logic, public health data, and machine learning to create a deployable risk scoring tool. Your work will form the foundation for in-app decision support.

Key Responsibilities

1. Research CKD Risk Factors

• Identify key clinical and demographic variables linked to CKD (e.g., age, diabetes, hypertension, urinalysis results)

• Review datasets and clinical guidelines to define features for prediction

• Categorize users into risk levels (e.g., low, moderate, high) or create a probability-based score

2. Build the Predictive Model

• Use public datasets (e.g. UCI CKD, NHANES, MIMIC-IV) or CareforWe-collected data

• Clean, engineer, and preprocess data (e.g., encoding, imputation, normalization)

• Train and test models like logistic regression, random forest, or gradient boosting

• Evaluate performance using metrics like ROC AUC, precision-recall, calibration, and validation splits

• Explain model outputs using SHAP/LIME and identify top predictors

• Assist in packaging the model for integration into a mobile app or web platform (e.g., via API or TF Lite)

What You’ll Need

• Proficiency in Python (Pandas, scikit-learn, matplotlib, etc.)

• Basic understanding of supervised learning and data preprocessing

• Ability to work independently to research clinical concepts and translate them into code

• Interest in digital health, medtech, or public health analytics

• Bonus: Familiarity with OCR, biomedical datasets, or mobile deployment