At iStop, I designed end-to-end machine learning workflows, from raw data preprocessing through model evaluation. I also compared classification models and tuned their performance.
I applied feature engineering and cross-validation to estimate model performance, then translated analytical results into business insights.
In my Heart Failure Mortality Prediction project, I evaluated classification models and used feature scaling and performance metrics to improve F1-score.
I also developed a credit risk prediction model and implemented an end-to-end ML pipeline covering preprocessing, model comparison, tuning, and serialization with Pickle.
