About Me

I build and study AI for the parts of healthcare that fall through the cracks, including conditions that disproportionately affect women and have historically been ignored.

That conviction did not come from a research paper. It came from living with endometriosis, PCOS, adenomyosis, and other comorbidities that went undiagnosed for decades. The same years I spent earning my Masters in Computer Science and building ML systems, I was also quietly navigating a body the medical system had given up trying to understand. Managing multiple conditions meant keeping track of more specialists, records, appointments, and follow-ups than any one person should have to hold in their head, and nothing connected them except me. That gap between what I was capable of professionally and what I was being told about my own health pushed me toward a question I could not let go of: how much of what patients experience, and women in particular, is simply not being measured, modeled, or coordinated well enough to help them?

That question led me to found HealthSteward. I started building it for myself, and the open-source version (v1 released in August 2026, now at v1.0.1) processes after-visit summaries, organizes records and action items, and prepares you for your next appointment. It is built as a local-first agentic system, and a paper describing the architecture, “HealthSteward: A Local-First Agentic Architecture for Patient-Side Care Coordination,” is under review at the IEEE Journal of Biomedical and Health Informatics. A consumer app is in development, and you can join the waitlist or learn more at healthstewardai.com.

Motherhood added a new dimension to that question. The postpartum experience makes visible just how much of what women go through gets quietly reclassified as personal rather than medical: matrescence, maternal mental health, the cognitive and emotional load of caregiving. It is not that these experiences are rare. It is that we have not built the infrastructure to study them properly.

Research

As a Visiting Scientist in The Jackson Laboratory's EndoRISE Program (ML Research), I work with endometriosis patient data in collaboration with a single-cell biology lab, focusing on patient stratification and relationships between inflammation markers and immune comorbidities. In parallel, I conduct independent ML research on maternal mental health. My position paper, The ML Community Must Operationalize and Evaluate Matrescence as a Construct in Its Own Right, was accepted to the ML4H 2026 Findings track.

Background

Before this, I spent 6+ years at Autodesk building and leading ML products at scale: recommendation engines, NLP platforms, and applied AI systems serving hundreds of thousands of daily users, most recently as a Principal ML Engineer. I also served as my organization's Privacy Champion, vetting AI and ML projects for data ethics, PII compliance, and open-source legal standards. Going from shipping production ML systems inside a large company to founding my own meant carrying that engineering and leadership experience into much less certain territory.

Community

Giving back has always been part of how I work. I have been a member of Asha for Education for nearly a decade, volunteering and fundraising for education and for the upliftment of women, children, and underserved communities. At Autodesk, I led the Girls Who Code program for two years, helped organize the TechWomen initiative, and served on the founding Impact Measurement Committee of the Autodesk Women Network, where I led data volunteers measuring the disparities the organization should target. From day one as a new grad, I also volunteered for pro bono consulting through the Autodesk Foundation, contributing ML and data expertise to nonprofits and startups working on workforce development and social impact.

Outside of research and building, Bharatanatyam, an Indian classical dance form, has been part of my life since the age of four. 25+ years of practicing it teaches you things no classroom does: patience, precision, showing up for something difficult day after day. There is more overlap between classical dance and research than people expect. Both tell stories through structure, both demand rigor, and both require you to hold complexity without losing the thread. I am currently three years into a formal degree in Kathak, a second classical dance form. Apparently I am drawn to long, demanding, deeply rewarding commitments.

I'm always interested in connecting with researchers, clinicians, founders, investors, and others working to make healthcare more coordinated and patient-centered. See what I'm building, or reach out.