Machine Learning
Customer Churn Prediction
A supervised-learning study comparing classification approaches for identifying customers who may leave a service.
Case study, Machine Learning
A live explainable healthcare risk platform: a LightGBM diabetes risk model trained on 70,692 health records (BRFSS), with patient‑level SHAP explanations and role‑based access for patients, doctors, and admins.
Context
Healthcare prediction work is not only about producing a score. Models need reproducible experiments, understandable feature contributions, and clear boundaries around what a prediction can and cannot say.
My role: Designed and built as Machine Learning Intern (Remote) at Ezitech Institute (Eziline Software House); maintained and improved since.
Implementation
The workflow organizes dataset-specific preprocessing and classifier benchmarking, rather than treating different clinical prediction tasks as one interchangeable problem.
MLflow provides a record of model runs so evaluation can be compared and revisited instead of relying on notebook state.
SHAP is used to inspect feature contributions. Explanations describe model behaviour; they are not clinical reasoning or proof of causality.
A FastAPI service connects model operations to a Next.js interface, with Supabase used for application data and supporting functionality.
Evidence
The points below summarize the public repository and owner-provided CV where noted. Read the source for implementation details and the evaluation setup before drawing further conclusions.
LightGBM diabetes risk model on a held-out test split (BRFSS, 70,692 records); the live model card documents the full evaluation.
Not the internship's 0.83. That was the best across seven datasets. This 0.8222 is the live model alone, on its held-out test split.
De-identified healthcare dataset domains described in project documentation.
Experiment tracking is part of the documented stack.
Boundaries
Continue exploring