Try a demo · Read a research case · Review the evaluation
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 flagship integration point for healthcare equity research. A synthesis layer and 12 modular evidence pipelines connect clinical and regulatory developments, patient populations, commercial adoption, reimbursement, financial disclosures and valuation. Core value: Turn a source-backed observation into an explicit model assumption and a testable investment view.
|
The platform combines 12 evidence modules, nine biomedical workflows, and an investment synthesis layer. Python utilities retrieve data, compare structured records and calculate scenarios; host-loaded skills guide reading and interpretation. The standalone repositories remain authoritative for their specialist scripts; bundled modules are versioned snapshots.
| Case | Decision question | Reproducible output |
|---|---|---|
| Biotech: HELIOS-B / vutrisiran | How does clinical evidence differ from a completed approval decision? | Approval-gate and commercial rNPV sensitivities. |
| Medtech: Intuitive Surgical | How do procedure growth and recurring economics affect revenue? | Platform-specific drivers and revenue sensitivities. |
| Managed care: medical costs | What does a medical-cost-ratio change imply for earnings? | Reconciled premiums and static earnings/EPS sensitivities. |
These are dated research demonstrations using real primary-source facts and explicit modeling assumptions. They are not current stock recommendations or fully underwritten price targets.
Source-backed semantic development set · Case verification ledger · Choose an installation path
flowchart TD
Q["Research question"] --> E["Primary evidence and source ledger"]
E --> A["Assumptions and uncertainty"]
A --> M["Models and sensitivities"]
E --> C["Source and accuracy checks"]
C --> M
M --> T["Thesis, catalyst and falsifier"]
T --> R["Monitor and re-underwrite"]
R --> E
Detailed source workflow · Maintenance and synchronization
|
From trial results to underwriting assumptions. Organize trial readouts and regulatory events around approval probability, label, timing and clinical differentiation.
|
Understand the economics of market access. Connect coverage and payment evidence to access, pricing assumptions and healthcare business economics.
|
|
Test demand against observable evidence. Use prescribing, utilization and launch proxies to challenge commercial assumptions and revenue trajectories.
|
Bring financial discipline to the healthcare thesis. Examine company filings, accounting signals and insider activity alongside the scientific and commercial evidence.
|
Explore all 12 evidence modules
| Research layer | Integrated module |
|---|---|
| Clinical and regulatory | clinical-catalysts |
| Publications, guidelines and expert signals | evidence-catalysts |
| Safety signals | fda-safety-signals |
| Prescription utilization | rx-utilization |
| Epidemiology and patient populations | epi-demand |
| Coverage and reimbursement | cms-reimbursement |
| Provider adoption | provider-adoption |
| Provider economics | provider-economics |
| Procedure and coding exposure | procedure-exposure |
| Patents, exclusivity and loss of exclusivity | ip-exclusivity |
| International access | global-access |
| Company filings and accounting | sec-forensics |
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.
AI supports research and education; outputs require source verification and accountable human review. Protect patient information, disclose material AI assistance and retain uncertainty. These tools are not validated for patient care or autonomous investment decisions.
Read the responsible AI policy — practical standards for evidence, disclosure, privacy, human oversight and corrections.
Working on healthcare investing, biomedical data or AI-enabled research? Project issues and pull requests are welcome.
Contribute a case, adapter or reproducible failure
Research and educational tools—not medical advice or investment recommendations.
Public project capabilities vary; inspect each repository's documentation and limitations.