-Developed a classification model for house documents using the pretrained EfficientNetV3 model and created its API.-Developed a service that extracts signatures and images from documents using an object recognition model based on the YOLOv8 model trained again, and verifies them against specified rules.-Developed a service that performs operations in a web project where it generates linear regression models with different combinations within a dataset and outputs those models that pass specific tests.-Developed a service that utilizes an object recognition model based on the DETR model to detect tables within images and extract information from tables in documents.

Worked as a data scientist at Fongogo, where I was responsible for the end-to-end development of the company's newly launched project. I developed machine learning models customized to the specific needs by focusing on the storage and analysis of data generated in crowdfunding campaigns.

Gathering public data from websites using Python through web scraping method and collecting data using the Selenium library.Organizing the collected data, analyzing it, and storing it in NoSQL MongoDB.Modeling the analyzed data.

Developed models capable of learning and decision-making based on natural language processing sentiment analysis of 1,600,000 labeled Twitter posts. I compared these models to select the most suitable one. Models used include Naive Bayes, SVM, Logistic Regression, etc.(Pandas, Wordcloud, Sklearn, Tf-idf vectorizer, Pyplot) were utilized.