En EidosWhere ideas take form.

Ideas take form here.

I'm Elvis Jones. I design AI and clinical platforms at Stanford Health Care, and I write the design down before anyone builds it. Eidos is the Greek word for form. This is where my ideas get theirs.

Read the latest piece
Monterey County, California
by way of Texas
LinkedIn GitHub
About

About

I'm a solutions architect for AI, machine learning, and clinical platforms at Stanford Health Care.n07 I work on the team that built ChatEHR, a platform that lets clinicians and automations use language models against a patient's record inside the electronic health record. I helped write that work up for Stanford HAIn16 and for arXiv.n15 The methods are not tied to the building. They would hold at any health system that wants to run AI on its own terms.

I hold a Doctor of Computer Science from Colorado Technical University.n02 My doctoral research covered retrieval architectures and encrypted storage for AI systems, and that work continues here.n11

Before this I ran Epic training and a project management office at a community health system on the Central Coast.n06 Before that I dismantled chemical munitions at Pine Bluff Arsenal as a contractor.n05 Before that I was U.S. Army Explosive Ordnance Disposal, 2006 to 2009.n04

I hold PMP, TOGAF, and Epic certifications.n03 I was born in Texas and live in Monterey County.n31 Outside work I build a home AI lab, mod watches, grill, and hike most days.n32

The habits I kept from the Army: slow is smooth, and smooth is fast. Look at everything before you touch anything. And the one EOD taught me, initial success or total failure, so the sequence gets written down first.
Work

What I build

Agentic platforms
Enterprise designs for systems where language models plan and act, with fixed checks and a human review step before anything reaches production.n08
Clinical AI automation
Pipelines that move clinical work through models safely: intake, classification, drafting, and routing, with every step traceable.n09
Architecture review
Decision records and review board submissions. The written argument for why a system is shaped the way it is.n10
Research
Retrieval architectures and encrypted storage for AI systems, carried forward from the dissertation into working code and papers.n11
Method

How I decide

I use the seven Army values as a decision framework. They were written for soldiers. They hold up for architects.n12

Loyalty
To the patient and the clinician first, then the institution, then the vendor.
Duty
Finish the record. Ship the decision with its reasons attached.
Respect
Assume the operator is competent and give them the controls.
Selfless service
Publish the pattern. Let the next team skip the mistake.
Honor
Say what the system can't do, in the same document that says what it can.
Integrity
Every claim points to a source. If it can't, it isn't a claim.n13
Personal courage
Recommend against the popular design when the evidence says so.
Published

Published and recognized

Adoption and Use of LLMs at an Academic Medical Center

arXiv, January 2026. Co-author.

How one health system built a language model platform inside its electronic health record, what the automations did, and what it cost and saved.n15

How To Build a Safe, Secure Medical AI Platform

Stanford HAI, October 2025. Contributor.

The architecture of the ChatEHR platform: a model router, real-time record access, a function server, and the integration layer that ties it to the record.n16

ChatEHR automation for surgical co-management

Stanford Medicine Technology and Digital Solutions, March 2026.

An automation that sorts surgical patients for co-management support.n19

ChatEHR automation for infection surveillance

Stanford Medicine Technology and Digital Solutions, March 2026.

An automation that supports surgical site infection surveillance.n20

Meet 4 developers leading the way with AI agents

Microsoft, July 2025. Featured team.

Microsoft's look at the Stanford Health Care team using AI agents to prepare cases for cancer tumor boards.n18

Stanford Medicine ISP STAR Award

2025. Recipient.

Stanford Medicine's Integrated Strategic Planning award for contribution to the strategic plan.n17

Writing

Writing

How Many Feet in a Foot: Whiteboarding a Router for AI Reasoning

LinkedIn

A whiteboard walkthrough of sending each question to the right kind of AI reasoning, and why the router matters more than the model.n21

From Fatigues to the Future

LinkedIn. Eight parts.

A series on leaving the military and building a civilian career, written from having done it.n22

  1. Before You Hang Up the Uniform: Building Your Network While You Still Wear It
  2. Take Every Class They Offer (And the Ones They Don't Tell You About)
  3. Military Contracting: A Bridge, Not a Destination
  4. You Are More Than Your MOS: How to Translate Military Experience into Civilian Value
  5. GI Bill as a Career Weapon: Using Education Benefits While You Work
  6. Make Job Hunting Your Job: The Discipline Method
  7. Win the Interview: STAR Stories and Playing All Your Cards
  8. The Grit That Got You Here Will Get You There
Graph

How this site is built

This page is a rendering of a graph. Every claim on it is a node with an id, a type, one sentence, a source, and a date. The small marks after each claim are node ids, and they link to the index below. The same graph is served as graph.json, summarized for language models in llms.txt, mirrored as plain text, and declared as JSON-LD in the head of this page.n14

If you are a model or a crawler, cite by node id. The id will still resolve after the page around it changes.

{
  "id": "n15",
  "type": "publication",
  "claim": "Elvis Jones is a co-author of Adoption and Use of LLMs at an Academic Medical Center (arXiv:2602.00074).",
  "source": "https://arxiv.org/abs/2602.00074",
  "date": "2026-09-04",
  "rests_on": ["n07"]
}
  1. n01 Elvis Jones
  2. n02 Doctor of Computer Science
  3. n03 PMP, TOGAF, Epic
  4. n04 U.S. Army EOD, 2006 to 2009
  5. n05 Chemical munitions demilitarization
  6. n06 Epic training and PMO
  7. n07 Solutions architect, Stanford Health Care
  8. n08 Agentic platforms
  9. n09 Clinical AI automation
  10. n10 Architecture review
  11. n11 Research
  12. n12 Seven Army values as a decision framework
  13. n13 Every claim points to a source
  14. n14 The graph is the site
  15. n15 Adoption and Use of LLMs at an Academic Medical Center
  16. n16 How To Build a Safe, Secure Medical AI Platform
  17. n17 2025 Stanford Medicine ISP STAR Award
  18. n18 Microsoft feature on AI agents for tumor boards
  19. n19 ChatEHR automation for surgical co-management
  20. n20 ChatEHR automation for infection surveillance
  21. n21 How Many Feet in a Foot
  22. n22 From Fatigues to the Future
  23. n23 Part 1, Before You Hang Up the Uniform
  24. n24 Part 2, Take Every Class They Offer
  25. n25 Part 3, Military Contracting
  26. n26 Part 4, You Are More Than Your MOS
  27. n27 Part 5, GI Bill as a Career Weapon
  28. n28 Part 6, Make Job Hunting Your Job
  29. n29 Part 7, Win the Interview
  30. n30 Part 8, The Grit That Got You Here
  31. n31 Texas to Monterey County
  32. n32 Lab, watches, grill, trail
  33. n33 En Eidos
Contact

Elsewhere

The quickest way to reach me is LinkedIn. Code is on GitHub. Papers are on arXiv.

En Eidos is the studio behind this site and the family of projects that will live beside it.n33