Gal Sapir

MD, PhD

Staff Research Scientist · Health AI

hi@galsapir.com · galsapir.com

Google Scholar / GitHub / Sparse Thoughts (blog) / LinkedIn

I work in health AI, developing foundation models and benchmarks. I enjoy breaking ambitious research questions into concrete experiments and figuring out what evidence would answer them. I want to help shape the relationship between AI and health, including the problems we ask models to solve and the environments where they learn.

Research & experience

Pheno.AI

2023–present

Staff Research Scientist2025–present

Previously Data Scientist2023–2025

Developed expertise in multimodal, longitudinal health data through hands-on research across the Human Phenotype Project (HPP), a deeply phenotyped cohort of approximately 28,000 enrolled participants.

Selected projects at Pheno.AI

Conceived and built a benchmark for health AIPhenoBench

  • Turned 90 clinical questions into reproducible tasks for comparing AI models; built the system for ingesting data, running evaluations, and analyzing results.
  • Evaluated LLMs and tabular foundation models, analyzing performance and failure modes.
  • Used by all Pheno data science teams to define tasks, compare models, and guide model selection.

Led the conception and development of a health agentHealth Agent

  • Defined clinical tasks and built core tools for a metabolic health agent that interprets participant data and provides grounded guidance.
  • Designed and built clinical evaluations for grounding, numerical accuracy, and clinical appropriateness.
  • The research became the basis for a health-agent system now in beta with HPP participants.

Developed foundation models for human healthGluFormer & HealthFormer · Second author on both

  • Co-led study conception, model development, validation, and publication.
  • GluFormer: identified external cohorts and designed studies of clinical value and generalization; the study tested transfer across 19 cohorts.
  • HealthFormer: shaped the multimodal research and evaluation agenda, including UK Biobank comparisons and clinical-trial simulations assessed against published intervention outcomes.

Led a research collaboration with Samsung ResearchGalaxy Health pilot

  • Defined the research scope, coordinated contributors, conducted research, and presented findings to Samsung Research.
  • Integrated smartwatch, CGM, dietary, and clinical data from approximately 200 participants to study health indicators and glucose prediction.

Selected publications

  1. PhenoBench: Mapping What a Deeply Phenotyped Human Cohort Can Tell Us

    Sapir G, Diament A, Wolf A, et al. Manuscript, 2026. First author.

  2. Grounding Health AI: Architecture and Evaluation of a Domain-Expert Metabolic Health Agent

    Diament A*, Sapir G*, Gorodetski M*, et al. medRxiv, 2026. *Equal contribution.

  3. Simulating clinical interventions with a generative multimodal model of human physiology

    Lutsker G, Sapir G, Merino J, et al. arXiv, 2026.

  4. A foundation model for continuous glucose monitoring data

    Lutsker G, Sapir G, Shilo S, et al. Nature, 2026.

  5. Use of continuous glucose monitoring to stratify individuals without diabetes

    Sugimoto H, Sapir G, Keshet A, Kuroda S. Communications Medicine, 2026.

  6. Pre-activity glycemic prediction prioritizes post-meal movement

    Shilo S, Sapir G, Lutsker G, et al. medRxiv, 2026.

  7. Impact of daily diet on sleep quality

    Shkolnik M, Sapir G, Shilo S, et al. Nature Health, 2026.

Full publication record → Google Scholar

Selected writing at Sparse Thoughts: Benchmarking is the new data activation · Curation all the way down.

Earlier experience

Lynx.MD

2022–2023

Software Engineer

Developed Python/Django software and streamlined AWS serverless deployment and CI/CD.

Hebrew University of Jerusalem

2015–2021

Graduate Research & Teaching

Developed hyperpolarized MR imaging and spectroscopy methods, with signal processing and analysis in MATLAB and Python. Mentored students and taught computational medicine, histology, and physiology to over 500 students.

Sheba Medical Center

2020–2021

Medical Intern

Clinical rotations in general surgery, internal medicine, and emergency medicine.

Education

Hebrew University–Hadassah Medical School

PhD
Hyperpolarized magnetic resonance research; supervisor: Prof. Katz-Brull
2021
MD
Medicine · Cum laude
2019
MSc
Biomedical Science · Magna cum laude
2017
BSc
Biomedical Sciences · Cum laude
2015