What To-Do: Machine Learning Powered To-Do List

Worked with a group of two other people to implemented a unique to-do list which recommends task, by comparing your active or expired tasks with other users. Implemented a naive recommender algorithm in Tensorflow.js to calculate embeddings using Universal Sentence Encoder and match the most relevant tasks using cosine similarity. Additionally, I implemented login authentication (including encryption and sessions support) using NodeJS and MySQL database. Optimized the website performance using advanced tools like browser caching, cookies, sessions, and cloud hosting. Worked with NodeJS backend and HTML, JS client to display the To-Do tasks with intuitive UI, timer, and added functionality.