I work on customizing frontier models for downstream applications at Amazon AGI. Before this, I spent ~14 months in Ads improving recommendations to drive shopper engagement and Amazon's retail business. Prior to Amazon, I spent 7.5 years at Lawrence Livermore National Laboratory as a Project Lead on foundational and applied ML problems.
Career
Applied Scientist
Amazon AGI
05/2025 – Present
Applied Scientist
Amazon Ads
01/2024 – 05/2025
Principal Scientist
Lawrence Livermore National Laboratory
10/2016 – 12/2023
Postdoc
IBM Research Almaden
04/2016 – 10/2016
PhD
Arizona State University
08/2010 – 03/2016
Recent Work
AWS Re:Invent 2025
Amazon Nova Forge
The easiest and most cost-effective way to build your own frontier model.
NeurIPS 2024
On the Use of Anchoring for Training Vision Models
Explores shortcomings of anchoring for ViTs and fixes them to show significant improvements in ImageNet-1K OOD detection, calibration, and generalization benchmarks.
ICML 2024
PAGER: Accurate Failure Characterization in Deep Regression Models
Failures in classifiers are clear (mis-classification rate) and well studied (OOD detection etc.), but very tricky in regression problems. PAGER takes a step in formalizing failure analysis using a combination of epistemic and non-conformity scores.
ICLR 2024
Accurate and Scalable Estimation of Epistemic Uncertainty for Graph Neural Networks
Introduces G-ΔUQ (pronounced "G-DUCK"), a novel training protocol for graph neural networks that supports reliable epistemic uncertainty estimates for graph classification and node classification tasks. Lot of SoTA results!
IOP ML Sci & Tech
Transformer-Powered Surrogates Close the ICF Simulation-Experiment Gap with Extremely Limited Data
New training and model selection protocol for Transformer surrogates, which improve existing methods on few-shot adaption to real world ICF experiments by ~40% improvement in predictive performance.