A service for preventing user outflow from a web resource using RL methods.

Analysis of the available data log, its processing, analysis of literature on the topic and methods used in business, creation of a model for recommending actions aimed at preventing user churn, creation of a service based on django for interaction with the created model.

Creating a model for recognizing pneumonia from X-ray images

Analysis of the found dataset, creation of CNN solution architecture, testing and validation of the obtained model.

Recommender system

Full project development, starting with analysis of available data, selection of architecture/models/metrics, implementation of the selected architecture, selection of hyperparameters, testing, optimization, implementation of the final solution in the form of py-scripts and a docker image.