Anthony G
- Research Program Mentor
MS candidate at Stanford University
Expertise
Computer Science. Software Development (C++, Python) Experience teaching intro math courses; doing research in CV video understanding and experience in NLP.
Bio
Hello, my name is Anthony and I am a grad student at Stanford. I did my undergrad at University of New Mexico with dual degree's in Computer Science and Applied Mathematics. While at UNM, I worked on C++ development and specialty projects at Sandia National Labs for 4 years. I have spent 2 internships on ML teams at Amazon and an internship at a fintech company. At Stanford, I have worked on projects for Object Detection/stable Object Tracking with stereo cameras, Sentiment Analysis for financial subreddits, and building new models for Child Speech Recognition. I am currently working on a pedestrian intent prediction research project in the Stanford Vision Lab. I am highly interested in any fields touching self-driving vehicles. In my free time, I like finding new ways to exercise, playing board/video games with friends, and spending time with my 2 dogs. In high school, I was a multi-federation national champion in powerlifting and still do powerlifting to this day. I am an avid SCUBA diver and am just a handful of courses away from being a Master Diver. I've recently picked up Brazilian jiu-jitsu and have fallen in love with the art and tough exercise.Project ideas
Value of Social Media for Stock Investing
Financial investing has evolved considerably in the last decade. Single tweets[1] can cause a stock to nosedive 10% in one day of trading. Can we harness these conversations on social media for stock investing? This project would involve hooking up to the relevant datasets' APIs and processing them with NLP algorithms. Some social media platforms will be more valuable for this than others. Exploring "high-value" social media targets vs. sentiment analysis across the platform could be interesting. [1] https://www.wsj.com/articles/tesla-stock-falls-after-ceo-tweets-stock-is-too-high-11588348672)
Can we detect every cell in the human body?
Our body is made up of trillions of cells, but what are they? Inspired by the Human Protein Atlas[1], we could potentially classify millions of different humans cells to better understand the human body and its' mechanisms. This project would involve using standard Computer Vision algorithms to segment and classify cells in a given data set. Performing error analysis and iteratively improving the algorithm would be part of the end-to-end ML pipeline. [1] https://www.proteinatlas.org/humanproteome/cell