A named
research
program.

Externally Governed Learning Systems (EGLS)

Confirmation 9951. Patent Center 74325060. Foundation of the research program and architectural core of the FxMED OS and FxMED Advisor products. Originated from Harvard GCSRT shortcut learning thesis work, published as SSRN 6160268 and Zenodo 10.5281/zenodo.18472595.

Cited as the underlying framework in the MedBoard AI follow on appendix and in the DrugSynthAI governance module.

MedBoard AI · Adaptive Learning Engine

Confirmation 6954. Patent Center 74992085. Three tier Analyst, Strategist, Judge machine learning architecture with EGLS governance. Covers attention adaptive study prescriptions, recursive self improvement, peer learning, clinical simulation, knowledge graph construction, and career outcome prediction.

Twenty seven claim follow on specification filed 2026 · 03 · 24. USPTO Trademark filed the same day (Serial 99721498, Class 042).

DrugSynthAI · Governance First Drug Discovery

Confirmation 9061. Patent Center 75043858. Three amendments filed (75043858, 75060417, 75093911). Total fifty one pages. Claim sets A through F. Protection of the multi agent pipeline architecture for de novo molecule design targeting genetically defined disease states.

Dual track: academic research record on SSRN, Zenodo, ChemRxiv, and Research Square; commercial protection through provisional and planned non provisional filings.

DrugSynthAI

Governance first AI pipeline for de novo molecule design targeting genetically defined mitochondrial defects. Sixty seven autonomous agents across five tiers (eleven pipeline agents, eleven validators, twelve orchestrators, five intelligence agents, eight precision medicine agents) plus fifteen assistants and five specialists. Five hundred thirty six passing tests including thirty seven security tests. One hundred sixty seven API endpoints.

Three validated discovery campaigns complete. One hundred sixty three candidates ranked per campaign. Eighteen stage gate decisions, all passed. Zero kill switch activations. Build discipline under numbered rules R01 to R103.

Revenue tiers: Explorer free, Precision Medicine at 299 dollars per month, Professional at 999 dollars per month, Enterprise on contract. Repository at github.com/fxmedus/drugsynth-ai. AIDD-GOV open standard v0.1 under Apache 2.0.

FxMED OS · Intelligent Physician Operating System

Clinical operating system for private practices with nineteen complete service layers: patient registry, scheduling, queue flow, encounter management, clinical documents, consent and permissions, attachments and media, portal policy, communications, AI scribe, tenancy, task workflow, entitlements, search, onboarding, white label, proxy access, membership, CRM, and guided setup.

Nine hundred fifteen passing backend tests. One hundred thirty seven passing frontend tests. Forty nine bounded contexts under a monolithic Python application. Four hundred forty canonical patients in production with zero test pollution. ES256 JWT authentication with HS256 legacy fallback.

AI Scribe dormant pending business associate agreement gate. Governance enforced through SOGDR ledger, SIRLP recursive learning protocol, Watcher Protocol, and a formal P0 to P2 fix register.

MedBoardPRO · Adaptive Licensure Preparation

Adaptive licensure preparation engine for USMLE and related medical boards. Three tier Analyst, Strategist, Judge machine learning architecture under the EGLS governance framework. Eighteen entity Prisma schema. Forty five plus API endpoints. Seventy eight documented requirements. Sixty one acceptance tests. Full OpenAPI specification.

Harvard inspired landing page. Gold accent design system. Shield logo v6. Brand protection across medboard.pro, medboardpro.com, and medboardpro.app domains. USPTO Trademark MedBoardPRO (Serial 99721498) filed in Class 042 under Section 1(b) Intent to Use.

FxMED Advisor

Clinician trained AI companion. Second production instantiation of externally governed learning systems in the FxMEDUS portfolio, alongside FxMED OS itself. Operates within the EGLS framework and is embedded inside FxMED OS as an operator specific AI layer, distinct from the generic documentation scribe.

Product thesis: inverse of generic large language model wrapped in medical prompts. The companion is trained on the operator specific workflow decisions, communication style, clinical reasoning, and practice preferences. Intellectual property strategy under evaluation: continuation under the EGLS non provisional or separate provisional.

AFMI Platform · USMLE Academy

Physician first educational SaaS platform. Technology stack: Next.js 15, FastAPI, Clerk v7 authentication, Supabase, and Stripe billing skeleton. Repository at github.com/fxmedus/afmi-platform.

Component extraction reactivated for FxMED OS integration. Extractable components include the Clerk JWT authentication guard pattern, the Clerk to database webhook upsert pattern, and the multi tenant membership model with tenant scoped request headers.