Spencer Austin Profile pic

Hi there, I'm

Spencer Austin

Currently: Columbia University

Get to know me

boat cruise headshot

Far from my hometowns of Chandler, AZ and Berkeley, CA I’m a fourth-year undergraduate at Columbia University studying analytics, computer science, and applied math. Lately, I've been interested in artificial intelligence/machine learning, techincal program management, and quantitative trading. I recently concluded a semester abroad at the University of Oxford where I studied Management and Philosophy. Keep scrolling to get to know more about me!

Some of my projects

Independent Research

Project 1 Image

A Hybrid LSTM-Transformer Neural Network for Enhanced Time-Series Prediction

I am synthesizing a hybrid neural network model combining LSTMs and Transformers for predicting individual stock price data using the TensorFlow, Keras, & scikit-learn libraries for Python. I have backtested and demonstrated improved accuracy and consistency over standalone counterparts using historical price data. I'm currently refining the model's parameters and training it on more data.

Industry Work

Project 2 Image

Custom FAQ Chatbot

Working with the UK-based fintech company Uome, I built a bespoke FAQ chatbot from scratch using Python, BERT, semantic search, and LangChain. With over 97% accuracy, the model guides the user to relevant resources based on queries posed in plain English. Although the AI model is not currently in use, building it was an exciting and challenging way to apply my knowledge to a real-world setting.

University Course Group Project

Project 3 Image

Chess Opening Repertoire Database

Watch my presentation of my and Ben Vazzana's social application where users can explore various chess openings and share games with friends. We built this for our Introduction to Databases Course using Python, PostgreSQL, HTML, JavaScript, and Flask. Our project was one of four to be recognized as an outstanding project, selected from a pool of 100+ projects.

High School Research Project

Project 4 Image

A Unique Quantum Approach to Finding the Maximum Number of Moves Required to Solve a Rubik’s Cube

For my high school senior research project, I proposed an approach to determine the minimum number of moves required to solve an nxnxn Rubik's Cube from any starting position using quantum computers. I worked with Jesse Nims and Xihong Peng, PhD.

Some of my interests

Traveling

Rubik's Cubing

Rubik's cube

Ever since I solved my first Rubik’s Cube in 4th grade, I’ve been obsessed with getting faster—my current record is 14.75 seconds! Along the way, I became fascinated by the combinatorics behind Rubik’s Cubes which inspired my research project on the calculation of graph diameter of an nxnxn Cube Cayley Graph (essentially the minimum number of moves needed to solve a Cube from any starting position) using quantum computers.

Running, Biking, & Hiking

When I'm not working, I'm usually biking through Central Park, running through Riverside Park, or wishing that there were more hiking trails near NYC. I most recently ran the Oxford Half Marathon in October 2024. Follow me on Strava!

Rubik's cube

Socials

Thanks for checking out my page! Keep in touch...