What I have built and run

The problem

Companies raising capital have their story spread across decks, financials, filings and public records, and the gaps between them are what investors find first. Checking that by hand is slow, and a generic AI summary is not trustworthy enough to act on.

What I built

RaiseForge turns a company’s documents, financial materials and public records into one verified record, then reads it the way an investor would. Its first market is life-science companies, but the pattern is industry-neutral: messy documents, regulated data and multi-step AI that has to be right before anyone acts on it. I own the architecture, data model, security model and test strategy.

• Two web applications, one for companies and one for investors, on a Postgres data model I designed.
• Tenant isolation enforced in the database with row-level security across the whole schema.
• A model gateway across AI providers, with per-call metering and a hard cost cap on every run.
• Document extraction from decks and filings, plus live connections to eight federal data sources: SEC EDGAR, NIH RePORTER, Grants.gov, SBIR.gov, FDA, ClinicalTrials.gov, USPTO and SAM.gov.
• A grounding layer: every AI finding has to be supported by its sources, and a separate verification pass has to agree before anyone sees it.
• An orchestrator for scheduled and multi-step runs, with every step logged.

Result

The investor read runs eight checks in parallel against a company’s documents and public filings. In tests on real companies it surfaced contradictions a reviewer would otherwise have missed, including a funding history that the company’s own SEC filings disproved. Each read runs for well under a dollar, under a hard cap.

• Next.js
• Postgres
• Row-level security
• LLM gateway
• Document extraction
• Federal data APIs
• Microsoft Graph
• Regression testing

The problem

A large health system wanted out of running its own analytics organization, but what it needed instead was not defined.

What I did

I built a managed analytics business that combined Optum’s analytics platforms, its care coordination platform and its consulting staff. I worked out what the health system’s analytics organization actually delivered and to whom, shaped a managed service and operating model that could take it over, and led the sale: a nine-figure, multi-year agreement in which the health system outsourced its entire analytics organization to Optum.

I then led the operational transition onto our platforms, keeping the reporting the business ran on working throughout.

Result

As COO/CFO I went on to run the operations and finance of that business, a $200M+ P&L.

I set up the enterprise-wide evaluation of AI and machine-learning investments across UnitedHealth Group, bringing hundreds of separate initiatives into one view of spend and value so leadership could see what was paying off.

I also ran an analytics and AI team that turned large claims datasets into contract decisions.

I served as interim CEO of a digital health company and have advised early-stage healthcare companies on operations, finance and growth. It is the same work as the interim and fractional COO engagement: step in, take ownership of the operation, and leave it running better than I found it.

Before the business side, I built software for about ten years, designing and building business systems in .NET, Java, COBOL and relational databases across financial services, information services, IT services and retail.

• Associated Financial Group, technical architect and team lead: designed business data solutions and led developers and data experts through every layer of the software life cycle, as the bridge between business lines and engineering.
• SUPERVALU, product manager for business systems: consolidated software, analytics and reporting for retail market development.
• The Thomson Corporation, senior software engineer and project lead.
• Unisys, programmer and systems analyst who became a mentor and team lead.
• U.S. Navy Reserve, information technology: software, database, network infrastructure and IT security projects.