This project focuses on developing a robust framework for processing continuous sensor data using change point detection, tokenization, and feature embedding. The goal is to encode medical sensor data into a language interpretable by machine learning models, particularly for use in NLP and time-series analysis.