🔥 Portfolio 🔥

	Recommendation System [RecSys] Creating recommendation system, based on users ratings, theirs comments and main information about board games from boardgamegeek.com 🛠️ Realised: data scraping, collaborative filtering, item-based recommendation, hybrid recommendation system.
	Image Classification [CV] Creating a multi-class classification computer vision model based on CNN. Based on a dataset containing footage from «The Simpsons» cartoon, training the model to predict the residents of Springfield in different years. And also test the model on test data in the Kaggle competition. 🛠️ Realised: image preparation, image augmentation, fine-tuning cv model.
	Multi-Class Classification [ML] Building a multi-class classification model based on decision trees. From a dataset describing the flags of different countries, the model predicts the official religion of the countries. 🛠️ Realised: feature engineering, feature selection, bagging, boosting.
	Object Detection [CV] Deploy computer vision model trained on COCO dataset using Streamlit. 🛠️ Realised: image preparation, image postprocessing, deploy web app.