Puja Trivedi

I am an Applied Scientist at Amazon, where I work on large-scale ML systems for product understanding and substitution. Before Amazon, I completed my PhD in Computer Science and Engineering at the University of Michigan, where I was advised by Prof. Danai Koutra in the Graph Exploration and Mining at Scale (GEMS) Lab.

My recent work focuses on large-scale applied machine learning and model reliability. I am particularly interested in building and improving ML systems through better data, model adaptation, and iterative development. More broadly, my background spans representation learning, graph ML, uncertainty, and large-scale modeling, and I am excited by problems that combine strong research taste with practical system building.

Email  /  CV  /  Google Scholar

We introduce GCLR, an active-learning framework for improving GNN-based graph clustering with LLM guidance.

We introduce G-ΔUQ, an accurate and scalable strategy for obtaining reliable uncertainty estimates for node classification and graph classification tasks.

We study how adaptation protocols can induce safe and effective generalization on downstream tasks through the lens of feature distortion and simplicity bias.

We rigorously study the effectiveness of popular scoring functions under distribution shifts and corruptions

We provide a novel generalization analysis for graph contrastive learning with popularly used, generic graph augmentations. Our analysis identifies several limitations in current self-supervised graph learning practices.

We contextualize the performance of several unsupervised graph representation learning methods with respect to inductive bias of GNNs and show significant improvements by using structured augmentations defined by task-relevance.