Avi Trost

Hi! I am a third year Computer Science PhD student at University of Wisconsin–Madison working with Fred Sala. As an undergraduate at Brown University, I was fortunate to work with Stephen Bach.

Last summer I did a research internship at KRAFTON AI.

My research interests include elastic architectures, efficient LLM inference, and data-centric ML. Feel free to reach out for a chat!

Education

Publications

COLM 2026

Test-Time Scaling Makes Overtraining Compute-Optimal

Nicholas Roberts, Sungjun Cho, Zhiqi Gao, Tzu-Heng Huang, Albert Wu, Gabriel Orlanski, Avi Trost, Kelly Buchanan, Aws Albarghouthi, Frederic Sala

Paper

Modern scaling laws.

AdaptFM Resource-Adaptive Foundation Model Inference @ ICML 2026

Fully Nested Transformers

Avi Trost, Alexander Yun, John Cooper, Gabriel Orlanski, Frederic Sala

PaperBlog

Nested models that share computation.

ICML 2026

Weight Updates as Activation Shifts: A Principled Framework for Steering

Dyah Adila, John Cooper, Alexander Yun, Avi Trost, Frederic Sala

Paper

Expressivity of steering.

Data in Generative Models @ ICML 2025

R&B: Domain Regrouping and Data Mixture Balancing for Efficient Foundation Model Training

Albert Ge, Tzu-Heng Huang, John Cooper, Avi Trost, Ziyi Chu, Satya Sai Srinath Namburi GNVV, Ziyang Cai, Kendall Park, Nicholas Roberts, Frederic Sala

Paper

Reclustering data and reusing training gradients for data mixing.

ACL Findings 2024

Learning to Generate Instruction Tuning Datasets for Zero-Shot Task Adaptation

Nihal V. Nayak, Yiyang Nan, Avi Trost, Stephen H. Bach

PaperCode

Training a model to generate conditional synthetic data for task adaptation.