Saurabh B
- Research Program Mentor
MS at University of Michigan - Ann Arbor
Expertise
Machine Learning, Data Science, Data Analytics
Bio
I recently completed my master's degree in Data Science at the esteemed University of Michigan and have accumulated three years of valuable professional experience in the fields of machine learning and data science. During this time, I had the privilege of working with renowned companies such as Twitter and Goldman Sachs, where I honed my skills in machine learning modeling and gained expertise in managing the entire lifecycle of models. In addition to my academic and professional pursuits, I have a genuine passion for exploring new destinations and embarking on thrilling adventures. Traveling to unfamiliar places allows me to immerse myself in diverse cultures, broaden my perspective, and create cherished memories along the way. Moreover, I find joy in socializing and making new connections, whether it's by enjoying vibrant parties or simply relaxing and spending quality time with newfound friends.Project ideas
Filtering Out Toxic Tweets
We are going to build a searchable Twitter feed that filters out “negative” content. Negative content will be tweets characterized by severe toxicity, insults, profanity, identity attacks, threats, and sexually explicit material. The goal of this project is to address a large part of what makes current social media platforms toxic for people of all ages: echo chambers of negativity, anger, and despair. The problem of negative content on social media feeds has not been solved. Social media platforms like Twitter, Facebook, Reddit, and more have not yet implemented any features that allow users to reduce the negatively characterized content with which they engage. Additionally, any form of a filter implemented by these companies could embroil them in more public relation debacles about free speech and censorship. Our project aims to address what these companies might be unable or unwilling to do: allowing users to opt out of unsavory content. An application like this will be useful for social media users that are exhausted by the incendiary language that plagues these sites today and would be a useful exercise in how a larger social media site might implement a version of this. This project will require data scraping methods (use of APIs), data storage and integration, modeling (Machine Learning techniques), implementation of information retrieval methods, and an end-user feature as a plugin, which will be a good exercise for learning end to end machine learning.