PhD Researcher · Human-Centered ML

Aria Mehta

I study how people and machine learning systems understand each other, building interfaces that are more interpretable. Currently a PhD candidate at the University of Edinburgh.

Currently seeking postdoc & research collaborations
AM
About

I'm a third-year PhD candidate in the School of Informatics, advised by Prof. Elena Vasquez. My work sits at the intersection of machine learning and human-computer interaction, where I design tools that let people interrogate and steer the models that increasingly shape their lives.

Before Edinburgh, I completed my MSc in Computer Science at the University of Toronto and spent a year as a research engineer working on clinical decision-support systems. When I'm not running experiments, you'll find me sketching type, hiking the Pentlands, or making an unreasonable amount of tea.

Research Interests
Interpretable ML Human-AI Interaction Algorithmic Fairness Visualization Model Auditing Trust & Calibration Mixed-Initiative Systems
Selected Publications
CHI 2026 · Best Paper Honorable Mention

Glassbox: Letting Practitioners Probe What Their Models Forgot

Aria Mehta, Daniel Okonkwo, Elena Vasquez

NeurIPS 2025

Calibrated Disagreement: Measuring When Humans Should Override AI

Priya Nair, Aria Mehta, Elena Vasquez

FAccT 2025

The Cost of a Second Opinion: Fairness in Human-in-the-Loop Triage

Aria Mehta, Marco Bianchi

Education & Experience
2023 — Present

PhD, Informatics — University of Edinburgh

Human-centered machine learning. Advised by Prof. Elena Vasquez.

2021 — 2022

Research Engineer — Meridian Health AI

Built interpretable clinical decision-support tools for ICU clinicians.

2019 — 2021

MSc Computer Science — University of Toronto

Thesis on visual explanations for sequence models. Graduated with distinction.

Let's talk

Open to collaborations, talks, and research questions.