Explore the flagship   ·   Workflows   ·   Research projects   ·   Sustainable finance work

Better evidence. Explicit assumptions. Falsifiable investment views.
AI should make research more inspectable—not make judgment less accountable.

I build open-source tools for healthcare investing, biotech research and AI-assisted fundamental analysis. My work connects biomedical evidence and public datasets to the questions that matter for an investment thesis: what changes, how it changes the model, and what would prove the thesis wrong.

• Research: Biotech, pharmaceuticals, medtech, life sciences tools and healthcare services.
• Evidence: Clinical trials, regulatory catalysts, utilization, reimbursement, provider economics and company filings.
• Methods: DCF, risk-adjusted NPV, scenarios, catalyst analysis and evidence-to-valuation workflows.

The platform brings together ClinicalTrials.gov, FDA, CMS, SEC EDGAR and other primary-source research workflows. It is the canonical integration point; the standalone project repositories below remain available during migration.

Discover → underwrite → decide and communicate → monitor and learn.

The workflow below comes from the Healthcare Equity Research Platform. It shows how evidence engines, valuation, thesis development and ongoing review connect.

flowchart TD
    subgraph DISC["1 · Discover"]
        ST["screen-themes"] --> IN["initiate<br/>scoping → memo → sub-sector layers"]
    end
    subgraph ENG["Evidence engines (co-installed plugins)"]
        E1["cms-reimbursement"]
        E2["clinical-catalysts"]
        E3["provider-adoption"]
        E4["procedure-exposure"]
    end
    FIN[("Financial layer<br/>filings & transcripts · Quartr / EDGAR")]
    subgraph UW["2 · Underwrite"]
        MV["model-valuation<br/>economic unit → formula → scenarios"] <--> TH["thesis<br/>variant perception · steel-man · pre-mortem"]
        ME["meetings-experts<br/>1-on-1s · expert calls, MNPI-safe"] --> TH
        INTL["international<br/>HTA · Japan · China · UK"] --> MV
        ESG["esg-stewardship"] --> TH
    end
    IN --> MV
    ENG -->|EVIDENCE BRIEFS| IV
    FIN --> IV
    LEDGER[("evidence ledger<br/>briefs from all five plugins")] --> IV
    EA["evidence-assembler agent<br/>runs the engines end-to-end for a ticker"] -.-> IV
    MV --> IV["investable-view capstone<br/>6 layers → scenario matrix → price-implied<br/>→ falsifiable underwriting statement"]
    TH --> IV
    subgraph DEC["3 · Decide & communicate"]
        CO["comms-compliance<br/>IC memo · MNPI scrub · disclosures"]
        PO["portfolio<br/>sizing · catalyst concentration · pairs"]
    end
    IV --> CO
    IV --> PO
    subgraph MON["4 · Monitor & learn"]
        EARN["earnings<br/>preview → live triage → post-print"] --> SD["sell-discipline<br/>scorecards · watchlists · post-mortems"]
    end
    PO --> EARN
    IV -->|falsifiers & early signals| SD
    SD -->|lessons → base rates & thresholds| TH

    classDef skill fill:#dbeafe,stroke:#2563eb,color:#111827
    classDef data fill:#dcfce7,stroke:#16a34a,color:#111827
    classDef agent fill:#fef3c7,stroke:#d97706,color:#111827,stroke-dasharray:5 5
    classDef brief fill:#fce7f3,stroke:#db2777,color:#111827
    classDef note fill:#f3f4f6,stroke:#6b7280,color:#111827
    classDef ext fill:#ede9fe,stroke:#7c3aed,color:#111827
    class ST,IN,MV,TH,ME,INTL,ESG,CO,PO,EARN,SD skill
    class FIN,LEDGER data
    class EA agent
    class IV brief
    class E1,E2,E3,E4 ext

Blue = skills · green = data sources and stores · amber (dashed) = agents · pink = evidence outputs · violet = suite handoffs.

Explore the source workflow and implementation →

_{Reproduced from the project's workflow documentation on 24 September 2026. Labels and connections are preserved; this profile copy does not automatically sync with future repository changes.}

Domain expertise: Biotech · Pharma · Medtech · Life Sciences Tools · Managed Care · Healthcare Services · Digital Health

Research methods: DCF · rNPV · Scenario Analysis · Catalyst Research · Evidence Synthesis

Primary sources first. Important numbers should retain their source, vintage and limitations.

Evidence is not the conclusion. Observations, assumptions, calculations and investment judgment should remain distinguishable.

Models must be challengeable. A thesis needs explicit sensitivities, disconfirming evidence and a clear account of what would change the view.

Human judgment stays in the loop. These AI agents, skills and research workflows are designed to support institutional-quality due diligence—not to replace analyst accountability or imply independent certification.

Working on healthcare investing, biomedical data or AI-enabled research? Project issues and pull requests are welcome.

Start a project discussion   ·   Explore my sustainable finance work

_{Research and educational tools—not medical advice or investment recommendations.
Public project capabilities vary; inspect each repository's documentation and limitations.}