Haya Bugshan — هيا بقشان

Hi! I’m Haya, a computational neuroscientist in training. I build models that read what the brain is doing from the signals it leaves behind, like a pupil changing size or a wave of EEG, and I work to hold AI systems accountable to the people they affect.

I’m an undergraduate at Brown University (class of 2028), concentrating in Computational Neuroscience and Applied Mathematics, and a 2026 Carney Quantitative Scholar. I work across the Perception, Action & Cognition Lab, the Watanabe Lab, and Brown’s Center for Technological Responsibility. I’m from Jeddah, Saudi Arabia, and I’m working toward a PhD in computational neuroscience and machine learning.

I came to Brown planning to study mathematics and philosophy. The questions stayed the same, about minds and what we can know about them; the tools changed. I also write about all of it on my Substack, Tarjama.

Research

Foundation models for pupil dynamics

In the Perception, Action & Cognition Lab, with Joo-Hyun Song and Jason Ritt, I built PupilFM, a pupillometry foundation model pretrained on 1,000+ hours of recordings from 2,000+ people. I’m asking whether it learns general pupil dynamics that reveal cognitive load, arousal and attention on tasks it has never seen. Featured in Brown News.

Causal masking: the model only ever sees the past and learns to predict what comes next. Then we probe its embeddings to see how well they predict cognitive states.

Presented at the Brown University Summer Research Symposium, 2026: Learning Task-General Signals of Cognition from Pupil Dynamics via a Pupillometry Foundation Model. H. Bugshan, T. Nguyen, J. Lader, J. Song, J. Ritt.

Kenyon cell diversity in the fruit fly brain

In the Kaun Lab, with MaryGrace Trousdell, I use machine learning to study Kenyon cells, the neurons at the heart of the fruit fly’s learning and memory center, and what makes their different types distinct. More to share once the work is published.

Zooming into a fruit fly’s brain: Kenyon cells sit in the mushroom bodies (yellow), the fly’s learning and memory center.

LLM accountability for reproductive health information

At CNTR, with the Re/Imagination Lab, I help test how ChatGPT, Claude and Gemini answer reproductive-health questions in English and Spanish across different legal environments, designing the prompts that surface where they fail. With Michelle L. Ding, Harini Suresh and Suresh Venkatasubramanian.

Mapping the tech ecosystem behind deepfake abuse

I’m a designer and developer on NCII-ECOSYSTEM.ORG, where my work is making CNTR’s research on the deepfake-abuse ecosystem, Ding et al. (2026, ACM EAAMO), understandable to anyone, not just researchers. I built the site, created and built its animated story, created the research page, and curated the sources in its evidence database. With Michelle L. Ding, Maja Mishevska, Harini Suresh and Suresh Venkatasubramanian.

From paper to public: I translated the research into an animated story and a searchable database that anyone can explore, no academic background needed.

Cite the website: Michelle L. Ding, Haya Bugshan, and Maja Mishevska. “NCII-ECOSYSTEM.ORG.” 2026. https://ncii-ecosystem.org

In the press: Brown News, CNTR News, Brown CS blog.

Other fun projects!

I write a Substack, Tarjama (ترجمة, “translation”), translating artificial intelligence, neuroscience, 3arabisms, womanhood, and the occasional laugh.
I’m Vice President of the AI Robotics Ethics Society at Brown, where I lead recruitment, coordinate our initiative teams, and bring in speakers and partners working across AI, robotics and ethics.
I build things outside the lab too, from vision-language model experiments to data storytelling. Check out my GitHub.

Contact

haya_bugshan [at] brown [dot] edu

GitHub LinkedIn Substack CV