darvin yi — director of ml, upwork · palo alto

frontier AI, shipped.

I build AI systems that move from research to reality. Currently leading ML at Upwork, focused on agentic AI. Former researcher at Stanford, UIC, Cruise, and Magic Leap — with work spanning medical imaging, computer vision, and autonomous systems.

palo alto, ca stanford phd · princeton bs
01

Agentic AI

Multi-step reasoning agents that take real-world action across tools, APIs, and workflows.

02

Foundational AIML

Core ML research — learning algorithms, generalization, and representation at scale.

03

Computer Vision

Detection, segmentation, and understanding of visual data across medical and real-world domains.

04

Research → Industry

Bridging frontier methods to shipped products. From paper to production.

app · ai

Daily Tarot

A daily AI-generated tarot reading. Pull cards, set your intention, get a personalized interpretation.

app · ai agent

Personal Assistant

An AI agent wired into calendar, tasks, and life context. My personal mission control in the browser.

site · academic

Research Portfolio

Full academic portfolio — publications, career arc, and research areas.

agentic ai · upwork · 2025

UpBench: A Dynamically Evolving Real-World Labor-Market Agentic Benchmark

D. Yi, T. Liu, M. Terzolo, L. Hasson, A. Sinha, P. Mendes, A. Rabinovich

new · 2025

mammography · jama network open · 2020

Evaluation of Combined AI and Radiologist Assessment to Interpret Screening Mammograms

T. Schaffter, D. Buist, C.I. Lee, ..., D. Yi, ..., G. Stolovitzky — DM DREAM Consortium

high impact

federated learning · jamia · npj digital medicine · 2018–2021

Distributed Deep Learning Without Sharing Patient Data

K. Chang, N. Balachandar, D. Yi, et al.

~500 citations

organ segmentation · nature scientific data · 2020

CT-ORG: A New Dataset for Multiple Organ Segmentation in Computed Tomography

B. Rister, D. Yi, K. Shivakumar, T. Nobashi, D.L. Rubin

167 citations

brain metastases · jmri · stanford aimi · 2019

Deep Learning Enables Automatic Detection and Segmentation of Brain Metastases on Multi-Sequence MRI

E. Grøvik, D. Yi, M. Iv, E. Tong, D.L. Rubin, G. Zaharchuk

public dataset
2025 → now

Director of Machine Learning

Upwork — Agentic AI, Palo Alto

2020 → 2025

Assistant Professor & Co-Executive Director, Ai-O

University of Illinois Chicago — Ophthalmology & Health Equity in ML

2023

Staff Applied Research Scientist

Cruise — Autonomous Vehicles, SF

2020 → 2023

Principal Computer Vision Scientist

Sirona Medical

2018 → 2020

Deep Learning Scientist, Human-Centered AI

Magic Leap — Sunnyvale

2014 → 2020

PhD, Biomedical Informatics

Stanford University

I'm a machine learning researcher and engineer interested in building AI that works in the real world — not just on benchmarks.

My career has moved between academia and industry: a Stanford PhD in biomedical informatics, research roles at Magic Leap, Sirona, and Cruise, five years directing an AI center at UIC, and now leading ML at Upwork where I focus on agentic AI systems.

Outside of work, I'm a photographer, a builder of small web things, and someone trying to live a little more intentionally.