Eric Martin

I build operational AI infrastructure: durable, accountable systems designed to hold up under real-world pressure.

The direction I'm building toward is modular intelligent systems for family caregivers doing elder care alone. There are millions of them, coordinating appointments across health systems, tracking medications across providers, translating clinical language into something actionable, mostly in their heads and mostly without help. That is a solvable problem, and I am building toward solving it.

At OCRC, I architect and run the AI-augmented operations layer behind tax efficiency and asset protection planning. I manage active client cases from intake through delivery and build the systems the firm runs on.

Selected systems:

• Workflow Automation Suite. The firm's master client template, now in its tenth major version over eight months. Handles entity formation, document assembly, validation, and bulk operations across the active roster.
• Document Heatmap. Context-aware tracking and verification engine for multi-entity client engagements. Identity-in-depth architecture, multi-tag support, automated batch operations.
• Bank Package Assembler. CRUT / Trust / LLC document packaging with execution-status enforcement, mandatory affidavits, and PDF watermarking.
• Client Onboarding. Full folder-structure generation with entity skeletons, manifest documents, and secure credential transfer.

Representative outcomes:

• ~95% reduction in new-client setup time
• ~90% reduction in manual formatting time per document revision
• 10+ major releases of the master client template shipped against the active roster over eight months

The mandate: bridge complex legal strategy and operational execution with zero data loss, strict compliance, and rapid turnaround on high-stakes cases.

Stack: Google Apps Script · Google Sheets · Multi-LLM (Claude · Gemini · GPT) · pdf-lib · Drive API · HIPAA-aligned client data handling · Multi-stakeholder coordination.

The substrate underneath Modular Care Systems and everything else I build: a persistent autonomous AI running on private hardware, designed for accountability over performance theater.

• ~100 self-monitoring operational components handling daily operations, server health, security audits, and long-term memory consolidation.
• A tiered intelligence layer: local inference for low-cost reflex monitoring; cloud reasoning for complex decisions. Operational inference costs reduced by 97%.
• Integrated scheduling, voice I/O, and cross-platform messaging for real-time AI-assisted communication and task management across Signal, calendar, email, and home automation.
• Governed by a Field Manual: an internal governance document with mechanical pre-commit enforcement, failure-class durability tracking, and per-component cadence audits. AI as accountable infrastructure, not as demo.

A working application of the sovereign AI substrate I've built: a caretaker-side system, on private hardware, that handles the parts of elder care that fall through the cracks of consumer apps.

It coordinates appointments across health systems. It maintains medication context when a prescription changes. It translates clinical changes into plain language before a family member has to ask. It carries the cognitive load of holding it all in one head, because one person shouldn't have to.

This is not a product. It is a working system that has quietly made one elder's life work better. The direction I am building toward is making this kind of infrastructure available to the family caregivers doing this work alone.

Built on: SMART on FHIR · OAuth 2.0 + PKCE · Consent-gated external actions · Per-action audit log.

Twenty years of client work, all of it word-of-mouth.

Before AI infrastructure, I built and ran a systems and design consultancy from 2005 to 2024. No advertising, no investor capital, no marketing budget. Just sustained referral from clients who trusted the work enough to send other people to it.

Representative work spanned HIPAA-aligned physician platforms, multi-division talent agency systems, and payment infrastructure for Class A commercial tenants. The common thread: high-stakes stakeholders, no room for failure, and no hand-holding after delivery.

The operational discipline (build it so it runs without you) is the same discipline I now apply to AI infrastructure.

AI Infrastructure: Claude · Gemini · OpenAI · Ollama (local inference) · Tiered model routing · Prompt caching · Agent orchestration · Custom governance doctrine.

Operations: Google Apps Script · Sheets · Financial workflow automation · HIPAA-aligned protocols.

Systems: PHP/MySQL · JavaScript · CRM architecture · Multi-party logistics · Design systems.

Senior roles bridging AI operations and regulated industries: healthcare systems, healthtech, financial services, insurance. Strong preference for organizations whose work touches eldercare, aging-in-place, or family caregiver coordination. Open to remote, hybrid, or Ohio-based.