Xiang Liu | Stony Brook University

About Me

Hi there! I’m an undergraduate student in the Computer Science Honors Program at Stony Brook University. Along my undergraduate time, I’m very fortunate to be advised by different professors to mentor and do research with me. Through my jobs & projects, my motivation is building new technologies that could benefit people around me. I always look to contribute in my future work.

Experience

August 2026 - Present
Undergradute Research Assistant · SPEL Lab
- Building a Vision Language Model to identify and geo-locate flooding events from images, videos and texts
- Collaborating with Department of Civil Engineering on NSF-funded project
- Advised by Professor Paola Cascante-Bonilla in Computer Science Department
- PyTorch
- Web Scraper
June 2026 - August 2026
Software Engineer Intern · Capital One
- Built a data lineage scoring system from scratch for Capital One’s Card Tech Data Engineering and Analytics Team
- Automatically traced how data moves between datasets, mapping upstream and downstream dependencies without manual input
- Scored each dataset on lineage accuracy, giving teams a clear signal of which datasets they could trust, saving manual Analysts’ review time
- Python
- TypeScript
- FastAPI
- React
- AWS
- Databricks
May 2025 - August 2025
Research Intern · Learning and Intelligence Lab
- Using image-generation models to generate action parameters accurately and efficiently for robots to execute home-tasks like putting books onto shelves without human intervention
- Presented at the April 2026 URECA Symposium, October 2025 MIT Undergraduate Technology Conference
- Advised by Professor Jorge-Mendez Mendez in the Electrical and Computer Engineering Department
- PyTorch
- Diffusion Models
- Flow Matching Models
- Robotics
- PyBullet Simulation
Poster
May 2024 - August 2024
REU Research Intern · SUNY Research Foundation
- Trained Deep Learning models like Fourier Neural Operators, U-Net and CNN for weather forecast in the United States
- Working towards fast and accurate weather forecasting without the computation from supercomputers
- Advised by Professor Yi Liu in the AI Research For Science Lab
- PyTorch
- Neural Networks
- Pands
- Xarray
Poster
December 2024 - Present
Lead Undergraduate Teaching Assistant · Department of Computer Science
- TA for CSE 214: Data Structures & Algorithms, and CSE 353: Machine Learning
- Teaching recitations, hosting office hours, making homework labs, and exam questions
- Enjoyed helping students and improving their grades to get admitted to the CS major

Projects

WorkIt

One place for both sides of hiring. Job seekers get fresh listings gathered and sorted from across the web, matched to what they are looking for, and a tool that fills in the applications for them. Businesses get a dashboard to manage applicants, track where each one is in the interview process, screen resumes, and see which candidates fit the role best. In progress.
Advising Bot

A chatbot that answers academic advising questions for Stony Brook students. It reads the entire undergraduate bulletin, so its answers come from the real requirements rather than guesswork, and it remembers what a student has already asked. A question that meant digging through the bulletin, or waiting on an advising appointment, now takes a few seconds.

Code
Car Model Brand Detector

A model that names the make and model of a car from a single photo. Built by training image recognition networks from scratch on a dataset of labelled car photos, then using Grad-CAM to highlight the parts of the picture the model actually looked at, so its answers can be checked rather than trusted blindly. A small project for the fun of pointing deep learning at something everyday.

Code
Molecule Optimization with LLMs and Genetic Algorithms

Drug discovery begins with finding molecules worth testing, and asking a language model for them alone tends to produce ones that could never exist. This project pairs the model with a genetic algorithm, a method that breeds and refines candidates over many rounds, so the suggestions stay chemically valid while still being new. It gets there without the cost of running several models against each other.

Report

Leadership

Peer Mentor

Outside of research, I’m quite involved on campus as a College of Engineering and Applied Sciences Peer Mentor where I mentor 3 first-year students each year to help them adjust to college life and “figure out what to do their college lives” along with 30 other mentors.
Resident Assistant

I am also a Resident Assistant (RA) at H Community where 1,300+ first-year students live in. I hosted large-scale events like Halloween Celebration called Werewolfie, Lunar New Year Celebration, Friendsgiving … that gained over 600 attendees with over $2k budget went into planning. My goal is to make college dorms home to students, and it has been a truly enjoyable and collaborative experience!