San Francisco · Physician · Clinical AI & Health Technology

Stephen G. Szeto, MD

Physician working at the intersection of clinical medicine and AI — board-certified physiatrist, Stanford Biodesign Innovation Fellow, and expert witness to the U.S. Senate on health technology.

Practicing clinician · 950+ research citations · Evaluating the technologies that promise to change care

Portrait of Stephen G. Szeto, MD

About

Clinician first, evidence always

I'm a physical medicine & rehabilitation physician trained at Stanford University, where I completed my residency and the Biodesign Innovation Fellowship — a year spent identifying unmet clinical needs and developing medical technologies to address them. Before medicine (MD, University of Toronto), I trained in mathematics and biology and earned a master's in medical science, and that quantitative footing still shapes how I work: I want the data, the methodology, and the failure modes, not just the demo.

Today I practice at North East Medical Services, a federally qualified health center in San Francisco, caring for an underserved older population at high risk of falls and disability. My clinical work spans musculoskeletal medicine, ultrasound-guided procedures, spasticity management, and electrodiagnostics — and I supervise internal medicine residents and lead monthly musculoskeletal ultrasound teaching. Frontline practice keeps my technology work honest: I evaluate tools against what actually happens in the exam room.

My throughline is a single question: does this technology actually make patients better — and how would we know? That question is about to matter more than ever, because AI is entering medicine faster than the evidence base behind it.

Focus

What I work on

Clinical AI & digital health

How AI and digital tools change clinical outcomes — from early work on AI in spinal cord injury rehabilitation to a widely cited evidence framework for mobile stroke rehabilitation. Peer reviewer for Frontiers in Digital Health.

Evaluation & real-world evidence

Systematic reviews and clinical research methods applied to emerging technology — currently leading a real-world evaluation of the first FDA-authorized wearable fall-injury mitigation device in older adults.

Policy & government advisory

Invited expert testimony to the U.S. Senate on technology and innovation in fall prevention. I translate between clinical evidence, technology teams, and policymakers.

Safety & health equity

What happens when technology meets patients with limited English, low digital literacy, and complex medical needs. Building culturally adapted, bilingual care programs at an FQHC — the deployment edge cases most products never see.

Public policy

Advising at the national level

U.S. SENATE SPECIAL COMMITTEE ON AGING · MAY 2026 · EXPERT TESTIMONY

Preventing Falls, Preserving Independence: Technology, Community Programs, and Innovation in Senior Safety

Invited by the U.S. Senate Special Committee on Aging to submit written expert testimony on innovation and technology in fall prevention for aging populations, entered into the Congressional Record and featured by the National PACE Association.

Research

Published across rehabilitation, digital health, and translational science

956citations
9peer-reviewed publications
2papers in top 1% most-cited
8h-index

J NEUROENGINEERING & REHABILITATION · 2023 · LEAD AUTHOR · TOP 10% MOST-CITED

Effect of mobile application types on stroke rehabilitation: a systematic review

A synthesis of 29 studies into an evidence hierarchy for how mobile technology supports stroke recovery — the kind of rigorous, clinically grounded evaluation that AI-enabled care tools need and rarely get.

Selected publications

Highlights

Full publication list on Google Scholar →

Looking ahead

Physicians belong in the room where AI gets built

The next chapter of medicine will be written by the teams building AI systems — and those teams need physicians embedded in the work, not consulted after the fact. My focus is clinical AI: designing evaluations that reflect real clinical practice, defining what safe model behavior looks like in medical contexts, and making sure these systems work for the patients who need them most, not just the easiest ones to reach.

I bring an unusual combination to that work: active frontline practice, formal training in needs-driven technology development, a research record in evaluating health technologies, and experience advising national policymakers. I'm open to clinical AI leadership roles — medical director, clinical AI lead, clinical informatics — with teams building at the frontier.

Beyond the clinic

Off hours

Basketball, fitness training, traveling with my partner, and keeping up with a Shiba Inu who does not respect the concept of a rest day.

Contact

Get in touch

Open to conversations about clinical AI, model evaluation in medicine, health technology, and how physicians can shape what gets built.

stephen@drszeto.ai LinkedIn Google Scholar