Alex T
- Research Program Mentor

MS at Columbia University

Expertise

Machine learning/ Artificial Intelligence, Big Data, Data Science, Signal Processing, Wireless Communications,

Bio

Provoking questions deserve equally satisfying answers. David Hilbert, one of the greatest mathematicians in history, believed behind every mathematical problem was a solution. "We must know, we shall know". He wasn't afraid to chase the big and "unsolvable" questions. One such problem led to his discovery of the Hilbert Transform, which provides the foundation of all modern communication systems. While not nearly as consequential as Hilbert, my graduate studies at Columbia University and professional experience have allowed me the opportunity to pursue challenging problems, and one of the great rewards of my life has been uncovering their answers. With that said, innovation is not limited to simply pairing solutions to problems, and can in fact take on many surprising forms! One of my favorite hobbies since childhood has been collecting and building Lego. Through the process of constructing a massive set like the UCS Millenium Falcon, brilliant build techniques and geometry work together to form a convincing replica of the iconic ship. In addition, I enjoy making my own derivative design changes to the recommended instructions. In the words of Han Solo - "I've added some special modifications myself".

Project ideas

Project ideas are meant to help inspire student thinking about their own project. Students are in the driver seat of their research and are free to use any or none of the ideas shared by their mentors.

Spotify Music Recommendation System

Create a Nearest Neighbor recommendation algorithm in Python to predict a playlist given an input song a user likes. This project would involve: • Harvesting a massive dataset of songs from Spotify via a developer account • Vectorizing and cleaning song metadata into feature library with Pandas DataFrames • Implementing ‘lazy’ learning algorithm for playlist generation • Experimentation with distance and similarity functions • Assess model performance by evaluating a test sample of users like their generated playlists

Coding skills

Python, MATLAB, C/C++, Algorithms

Teaching experience

TA for undergraduate linear algebra for two years, teaching lectures to 100s of students on Fridays. Formal mentor to a current Binghamton University student, through the Watson scholar's program. Designed and taught internal Lockheed Martin course on Reinforcement Learning with support from Columbia University faculty.