Founder · Scientist · Builder

I’m Han.
CTO & Co-founder
of Yumi.

I build AI that works with clinicians, grounded in the patient record and connected to the work of care.

At Yumi Health, I co-founded an AI-native platform for skilled nursing and lead its technical development. My work as Chief AI Scientist at Layr brings the same focus to clinical intelligence and MDS agents.

Building Yumi Connect on LinkedIn

Han Wang wearing a UC Berkeley sweatshirt — Han WangHealthcare AI · UC Berkeley

Clinical AI agentsHome-built retrievalHuman–AI interactionMulti-agent systems

Building in healthcare

From clinical records
to useful intelligence.

My goal is to advance the state of the art in MDS agents. The engineering starts with the evidence a clinician needs to make a decision.

01 / Yumi Health

Co-founding Yumi.
Building around care.

CTO & Co-founder

From building with clinicians on site to shaping the platform and founding engineering team. An AI-native record, a clinical copilot, and MDS work connected in one place.

Our home-built retrieval pipeline connects assessment questions to chart records, relevant dates, and evidence a clinician can inspect.

Read the Yumi story

Yumi / Connected clinical context

NotesOrdersAssessmentsResults

Home-built retrievalThe right record.
The right clinical context.Question relevance · Observation period · Source links

Evidence for reviewAnswer + explanation + supporting record

Clinician

A conceptual view of the retrieval workflow

02 / Layr

An MDS agent that
shows its work.

Chief AI Scientist

Building the retrieval and reasoning systems behind clinical assessment recommendations, with a path from each suggested answer back to its chart evidence.

A home-built pipeline combines question-specific search, BM25 relevance, source weighting, and assessment date windows to assemble focused clinical context.

Explore the Layr pipeline

Layr / Inside the MDS agent