Work

Everything I've shipped.

Systems that reached real users on real data — plus the one currently in development. Ordered by recency, not by how well they photograph.

$81M
Returned to the Medicare Trust Fund
10M+
Medicare & Medicaid lives monitored
2M+
Predictions served per day, at peak
Best in KLAS, healthcare AI/ML, 2021–2023

Measured at ClosedLoop.ai, 2021–2024

Bill & Melinda Gates Foundation, via Alithi  /  Higher education  /  2025–

Lucy, a strategy advisor for postsecondary transformation

College and university leaders mostly know what works — course redesign, proactive advising, credit mobility. What they don't have is a way to figure out which of it applies to their institution, at their stage, with their data. Lucy is the multi‑agent system that closes that gap.

It matches an institution against 7,000+ NCES records — Carnegie class, HBCU and Tribal designation — infers where that institution sits on a five‑stage transformation journey from behavioral signals, and returns only the research that fits, written to the reader's role and target population. Built on Google ADK and Azure OpenAI over an A2A protocol, and delivered as a public good.

Multi‑agentGoogle ADKPublic good
Health Universe × i‑Cubed  /  Clinical trials  /  2024–2026

Project LOOM

A documentation system that drafts a trial's entire document set as one coherent body of work: protocol, manual of procedures, safety plan, even patient recruitment materials. Nothing gets invented, and nothing drifts — edit one document and it surfaces every place that change contradicts the others. Cross‑document consistency is the hard problem here; generating any single document is not.

An i‑Cubed Center initiative, built with engineering support from Health Universe's dev team and validated and tested by Duke Clinical Research Institute.

AI architectAgenticCross‑document consistency
Health Universe  /  Oncology  /  2024–2026

The whole chart, read and sorted

A system that ingests a patient's complete medical record, determines which cancer it's looking at — patients often have more than one — and assembles a full clinical history for that specific disease. The summarizing isn't the hard part. Disambiguating which findings, treatments, and dates belong to which primary is.

Record ingestionMedical NERClinical histories
Health Universe  /  Revenue cycle  /  2024–2026

Getting pathology invoices paid

An agent‑to‑agent system that submits, resubmits, and appeals pathology invoices to payors. Before anything goes out it reviews each test against the justification on file and flags what's missing or too thin — the gaps that turn into denials. It pulls the patient's history to find the evidence that supports the test and surfaces it, so it can either be attached or sent back to the physician with a note naming exactly what's absent.

Agent‑to‑agentPayor workflowsDenial prevention
Babson Diagnostics  /  Clinical diagnostics  /  2022–2024

Counting platelets that don't want to be counted

Prick a finger and platelets start clumping at the wound within seconds, and a clumped sample reads as a falsely low count. It's the reason small‑volume blood counts were long considered unreliable — published studies still find platelet bias beyond allowable limits in roughly a third of capillary samples, which is a hard floor on how small a blood draw can get.

Babson's answer was the inverse of Theranos's — not miniaturizing the lab, but stabilizing a sample a tenth the size of a normal draw and running it on proven Siemens analyzers in a real one. That only works if the numbers coming off those analyzers survive the smaller sample.

I developed the algorithmic assay that made the platelet count hold up, and operationalized it from research prototype through commercial launch — then built the data operations behind that launch: order tracking, sample lifecycle, and financial monitoring at 1,000+ samples a day, on pipelines linking Sysmex hematology analyzers to the LIS and Salesforce.

Prototype → launchReal‑time MLInstrument integration
ClosedLoop.ai  /  Healthcare AI  /  2021–2024

$81M returned to the Medicare Trust Fund

Led the data science organization — 10+ data scientists, engineers, and project managers — reporting to the CEO as a non‑voting member of the executive team, and built the startup's professional services arm into a profitable line of business.

The models: chronic disease progression and hospital admission risk across 10M+ Medicare and Medicaid lives, serving 2M+ predictions a day on AWS — presented at HFSA and the ADA and published in the Journal of Cardiac Failure. The preventative programs they powered were recognized by NAACOS and the Wall Street Journal, and the platform took Best in KLAS for healthcare AI/ML three years running — above Epic and Cerner.

Team of 10+P&L ownershipBest in KLAS ×3

Also active

  • Sylvan Labs — applied AI advisory since 2026, now hands‑on build work
  • Atlas Mineral — B2B SaaS for precious metal order reporting and tracking, built for Alpax and running $1M+ monthly

Past engagements

  • SpyCloud — supply chain attack vulnerability
  • Capital Pulse — financial risk and forecasting
  • Beheld — analytics and workflow design
  • Perfaware — backend systems for pharmacy curbside pickup, 2020–2021

Start here

Most AI stops at the demo.

I take on a small number of engagements at a time, so I can be straight with you about whether yours is one I should take. Bring me the one that stalled.