At Technopark, I worked with a team developing a product recommendation system for store customers, using their previous purchases and expressed product interest. I prepared customer purchase and browsing data, engineered features, and trained the recommendation model.
For my Google Analytics Customer Conversion Prediction project, I built an end-to-end classification pipeline using 1.2 million user sessions and 4 million events. I compared CatBoost, Random Forest, and Logistic Regression models; the final CatBoost model reached 0.75 ROC-AUC and 0.70 accuracy.
I also saved the model and implemented a FastAPI REST API with a /predict endpoint. Before my Machine Learning Engineer role, I worked as a Sales Consultant at Technopark, where I ranked among the top-performing consultants and trained and supported new employees.
