Eric Xue

I'm Eric Xue, a master's student in computer science at Columbia University, with a bachelor's from the University of Toronto. I leverage machine learning and generative AI to build practical applications, from full-stack platforms powered by LLMs to vertical LLM agents for real-world tasks.

I've published research at AAAI and TMLR, and previously conducted research at the DREAM Lab (UIUC), Computational Social Science Lab (UofT), and Social AI Research Group (UofT).

Email  /  Resume  /  Twitter  /  Github

Publications

Proposed a curvature regularization technique, informed by theoretical insights, to enhance the inherent adversarial robustness of models trained on condensed datasets.

A personalized AI behavior model for chess that uses two-stage fine-tuning and meta-networks to efficiently predict individual decision-making styles with minimal data, enabling flexible human-AI collaboration.

An LLM agent framework that autonomously builds and iteratively optimizes image classification pipelines, including data augmentation, model architecture, and hyperparameters, using real training feedback without human supervision.

A hybrid machine Theory of Mind approach combining LLM-generated hypotheses with Bayesian inverse planning to predict mental states, enhancing performance on ToM tasks and enabling socially intelligent agent models.

Granular performance analysis using ROC curves of pre-trained object detectors in the autonomous vehicle setting, examining different data subgroups (bounding box sizes, object types, occlusion levels) and various image perturbations.

Selected Projects

A web platform that leverages an LLM pipeline to analyze students' experiences and interests, research target schools, and use a fine-tuned model trained on scraped and cleaned high-quality essay samples to help draft and improve college application essays. Earned over $1,000 in revenue from paying users.

Miscellanea

Awards