# Kevin D. Johnson

> HPC+AI Infrastructure, Neuromorphic Computing, Quantum-Centric Supercomputing, Forward-Deployed Engineering. Field CTO at IBM working on the HPC Cloud portfolio (IBM Spectrum Symphony, Storage Scale, and LSF), writing and building independently on neuromorphic computing, heterogeneous orchestration, and the infrastructure underneath large-scale AI.

This file indexes the machine-readable form of https://kevindjohnson.org. Every file below is generated from the published site, so it does not diverge from what a reader sees. Content is authored by Kevin D. Johnson; cite with attribution and a link.

## Start here

- [Full corpus](https://kevindjohnson.org/llms-full.txt): profile, all 77 published demonstration records, and all 27 articles in one file (~103,625 words).
- [Demonstration index](https://kevindjohnson.org/demos-index.md): all 77 records, one line each, about 20 KB. **Start here for anything about the demonstrations**, then open the one you want.
- **One page per demonstration** at `https://kevindjohnson.org/demos/<id>.md`, a few kilobytes each, carrying the full summary, the platform list, the announcement, the video, and links onward. Demonstration 81 is at https://kevindjohnson.org/demos/81.md.
- [Demonstrations as JSON](https://kevindjohnson.org/demos.json): every record in one structured file, for programmatic use. It is 123 KB, so a fetch tool with a size limit will cut it off partway; the index and the per-demo pages exist so nothing depends on that.

## Profile

I work at the intersection of high-performance computing, AI infrastructure, and the data platforms that hold them together. Most of my career has been spent building and deploying these systems for organizations where downtime or bad answers aren't an option. That includes financial services, government, defense, healthcare, and research.

I'm a Field CTO at IBM, focused on the HPC Cloud portfolio, which includes Spectrum Symphony, Storage Scale, and LSF. Before that, I spent over a decade in consulting engagements across the same stack. That work covered petabyte-scale parallel file systems, real-time workload orchestration, and the infrastructure underneath large-scale AI. Before IBM, I managed HPC biocomputing infrastructure at TGen, ran a specialty coffee company, and worked in storage virtualization and disaster recovery.

Lately, most of my independent work has been exploring what happens when neuromorphic hardware, large language models, and distributed orchestration frameworks are combined in ways their designers didn't anticipate. The articles on this site come from that exploration.

I hold an MBA and a Master of Accounting in Finance from Keller Graduate School, an M.A. in Theology from Fuller Theological Seminary, and an M.S. in Global Technology and Development from Arizona State University, where I'm currently a Ph.D. candidate in Innovation in Global Development. My research examines how coordination infrastructure shapes what we can know, decide, and imagine, using computational AI and simulation methods to compare architectures at scale, arguing that infrastructure building is itself a mode of development theorizing.

### Recent focus areas

- **Palantir Foundry / AIP** — Three integrations — entity discovery + ontology enrichment, multi-modal targeting, and autonomous ontology generation via SymWisdom.
- **LLM & Gen AI Infrastructure** — vLLM, GPFS KV cache sharing, semantic routing, Nemotron-120B reflection engine, multi-LLM orchestration at scale.
- **Neuromorphic & Heterogeneous Compute** — 10-chip Akida hive mind across AKD1000 and AKD1500, QPU+NPU+GPU+CPU orchestration, sub-millisecond edge inference across 12 silicon architectures.
- **Autonomous Ontology Design** — Systems that discover, name, and structure concepts from continuous perception — no human modeling required.
- **Data Platform Architecture** — Petabyte-scale GPFS, mmap'd shared state across 11 nodes, experience lifecycle management, three-tier KV cache replacing Redis.
- **Forward-Deployed Engineering** — 13 years delivering to government, research, financial services, healthcare, and life sciences.

## Research

The demonstrations are the instrument of a doctoral thesis, not a portfolio. The Ph.D. work at Arizona State asks how coordination infrastructure — the scheduling, placement, and storage layer that decides what runs where — shapes what a system can know, decide, and imagine. The argument is that building that infrastructure is itself a mode of development theorizing: an architecture encodes a claim about how coordination ought to work, and running it at scale is how the claim gets tested. That frame is what makes the demonstrations and the papers one body of work rather than a hobby and a reading list.

What that means if you are reading this as a developer:

- **The stack is held constant on purpose.** Akida for inference, Symphony for placement, GPFS for shared state. Fixing the substrate and varying the domain is what makes 77 demonstrations comparable to each other instead of 77 unrelated projects.
- **Each demonstration is an experiment with a stated result.** On the 13 records from demonstration 69 onward, the Findings list on a per-demo page is the measured outcome — latencies, accuracies, separations, agreement with simulation — kept separate from the Platforms list so you can filter on hardware without wading through results. On the 64 earlier records there is no separate Findings list and the measured result sits in the summary, so read the summary rather than concluding the demonstration reported nothing.
- **The domains are deliberately far apart.** Renaissance counterpoint, counterparty credit risk, cognitive electronic warfare, sleep staging. A coordination claim that only holds in one domain is a domain trick, not an architecture, so the spread is the evidence.

## Demonstrations

The demonstration program runs to 81 builds. 77 of them are published as records here, numbered 5 through 81 and dated 2025-12-29 to 2026-09-14, on the BrainChip Akida neuromorphic platform and a compute ontology stack of IBM Spectrum Symphony, LSF, and Storage Scale (GPFS). They span defense and national security, financial services, healthcare, critical infrastructure, information integrity, and AI infrastructure.

- [demos-index.md](https://kevindjohnson.org/demos-index.md) — complete, one line per demo
- `https://kevindjohnson.org/demos/<id>.md` — one small page per demonstration
- [demos.json](https://kevindjohnson.org/demos.json) — every record as structured data
- [demos.md](https://kevindjohnson.org/demos.md) — every record as prose

## Articles

- [Neuromorphic Compute: What Brain-Inspired Computing Really Means](https://kevindjohnson.org/neuromorphic-compute-what-brain-inspired-computing-really-means/): 2026-09-11, 3,161 words. Markdown: https://kevindjohnson.org/articles/neuromorphic-compute-what-brain-inspired-computing-really-means.md
- [Nothing Was Added](https://kevindjohnson.org/nothing-was-added/): 2026-09-09, 1,947 words. Markdown: https://kevindjohnson.org/articles/nothing-was-added.md
- [Down in the River: Sovereignty, Homogeneity, and the Floor Nobody Owns](https://kevindjohnson.org/down-in-the-river-sovereignty-homogeneity-and-the-floor-nobody-owns/): 2026-08-24, 2,480 words. Markdown: https://kevindjohnson.org/articles/down-in-the-river-sovereignty-homogeneity-and-the-floor-nobody-owns.md
- [Nothing Downstream Changed](https://kevindjohnson.org/nothing-downstream-changed/): 2026-07-30, 6,720 words. Markdown: https://kevindjohnson.org/articles/nothing-downstream-changed.md
- [Nobody Asks Whether a Bridge Corresponds](https://kevindjohnson.org/nobody-asks-whether-a-bridge-corresponds/): 2026-07-27, 5,996 words. Markdown: https://kevindjohnson.org/articles/nobody-asks-whether-a-bridge-corresponds.md
- [No One Builds Cathedrals by Judging Bricks](https://kevindjohnson.org/no-one-builds-cathedrals-by-judging-bricks/): 2026-07-24, 4,686 words. Markdown: https://kevindjohnson.org/articles/no-one-builds-cathedrals-by-judging-bricks.md
- [Old Black Water Keep On Rollin' - The Neuromorphic in a Heterogeneous Compute Ontology](https://kevindjohnson.org/old-black-water-keep-on-rollin-the-neuromorphic-in-a-heterogeneous-compute-ontology/): 2026-06-22, 3,104 words. Markdown: https://kevindjohnson.org/articles/old-black-water-keep-on-rollin-the-neuromorphic-in-a-heterogeneous-compute-ontology.md
- [Velocity Decoupled: The Architecture of a Neuromorphic Hive Mind](https://kevindjohnson.org/velocity-decoupled-the-architecture-of-a-neuromorphic-hive-mind/): 2026-05-17, 466 words. Markdown: https://kevindjohnson.org/articles/velocity-decoupled-the-architecture-of-a-neuromorphic-hive-mind.md
- [Where Capacity Really Lives and What Profiles Don't Say](https://kevindjohnson.org/where-capacity-really-lives-what-profiles-dont-say/): 2026-04-25, 2,707 words. Markdown: https://kevindjohnson.org/articles/where-capacity-really-lives-what-profiles-dont-say.md
- [Neither the Stack Nor the Swarm](https://kevindjohnson.org/neither-the-stack-nor-the-swarm/): 2026-04-17, 3,388 words. Markdown: https://kevindjohnson.org/articles/neither-the-stack-nor-the-swarm.md
- [Build It Anyway: Sankar and Karp on Institutional Decay](https://kevindjohnson.org/build-it-anyway-sankar-and-karp-on-institutional-decay/): 2026-03-30, 4,314 words. Markdown: https://kevindjohnson.org/articles/build-it-anyway-sankar-and-karp-on-institutional-decay.md
- [The Rickhouse on the River](https://kevindjohnson.org/the-rickhouse-on-the-river/): 2026-03-27, 2,293 words. Markdown: https://kevindjohnson.org/articles/the-rickhouse-on-the-river.md
- [SymPalantir RAG: Why RAG Security Requires Heterogeneous Compute](https://kevindjohnson.org/sympalantir-rag-why-rag-security-requires-heterogeneous-compute/): 2026-03-27, 2,726 words. Markdown: https://kevindjohnson.org/articles/sympalantir-rag-why-rag-security-requires-heterogeneous-compute.md
- [Just Still](https://kevindjohnson.org/just-still/): 2026-03-24, 4,702 words. Markdown: https://kevindjohnson.org/articles/just-still.md
- [Still Signal](https://kevindjohnson.org/still-signal/): 2026-03-23, 3,473 words. Markdown: https://kevindjohnson.org/articles/still-signal.md
- [The System Is Still Looking: Barrels, Encryption, and the Confidence Ceiling](https://kevindjohnson.org/the-system-is-still-looking-barrels-encryption-and-the-confidence-ceiling/): 2026-03-23, 4,685 words. Markdown: https://kevindjohnson.org/articles/the-system-is-still-looking-barrels-encryption-and-the-confidence-ceiling.md
- [Nobody Taught It to Look: How a Neuromorphic Hive Mind Builds Its Own Ontology](https://kevindjohnson.org/nobody-taught-it-to-look-how-a-neuromorphic-hive-mind-builds-its-own-ontology/): 2026-03-20, 3,528 words. Markdown: https://kevindjohnson.org/articles/nobody-taught-it-to-look-how-a-neuromorphic-hive-mind-builds-its-own-ontology.md
- [Intelligence Doesn't Start Anywhere and That's the Point](https://kevindjohnson.org/intelligence-doesnt-start-anywhere-and-thats-the-point/): 2026-03-13, 5,206 words. Markdown: https://kevindjohnson.org/articles/intelligence-doesnt-start-anywhere-and-thats-the-point.md
- [What the University Was, What It Wasn't, and What Comes Next](https://kevindjohnson.org/what-the-university-was-what-it-wasnt-and-what-comes-next/): 2026-03-06, 7,006 words. Markdown: https://kevindjohnson.org/articles/what-the-university-was-what-it-wasnt-and-what-comes-next.md
- [Executive Decisions at the Edge: Non-Extractive Targeting Built in Five Days with Symphony, Akida, and Foundry](https://kevindjohnson.org/executive-decisions-at-the-edge-non-extractive-targeting-built-in-five-days-with/): 2026-03-05, 2,462 words. Markdown: https://kevindjohnson.org/articles/executive-decisions-at-the-edge-non-extractive-targeting-built-in-five-days-with.md
- [The LLM That Cried Wolf](https://kevindjohnson.org/the-llm-that-cried-wolf/): 2026-02-27, 3,820 words. Markdown: https://kevindjohnson.org/articles/the-llm-that-cried-wolf.md
- [Developers in the Hands of an Angry Dave Ramsey](https://kevindjohnson.org/developers-in-the-hands-of-an-angry-dave-ramsey/): 2026-02-20, 2,270 words. Markdown: https://kevindjohnson.org/articles/developers-in-the-hands-of-an-angry-dave-ramsey.md
- [The Modernization Problem: What We Lose When We Stop Asking Questions About AI and Infrastructure](https://kevindjohnson.org/the-modernization-problem-what-we-lose-when-we-stop-asking-questions-about-ai-an/): 2026-02-13, 2,641 words. Markdown: https://kevindjohnson.org/articles/the-modernization-problem-what-we-lose-when-we-stop-asking-questions-about-ai-an.md
- [The 20% Question: Why AI Isn't Failing and What's Missing](https://kevindjohnson.org/the-20-question-why-ai-isnt-failing-and-whats-missing/): 2026-02-06, 1,851 words. Markdown: https://kevindjohnson.org/articles/the-20-question-why-ai-isnt-failing-and-whats-missing.md
- [The Nature of AI: Beyond Security Toward Partnership](https://kevindjohnson.org/the-nature-of-ai-beyond-security-toward-partnership/): 2026-01-29, 3,174 words. Markdown: https://kevindjohnson.org/articles/the-nature-of-ai-beyond-security-toward-partnership.md
- [Context-Driven Terraform for IBM Spectrum Symphony on IBM Cloud](https://kevindjohnson.org/context-driven-terraform-for-ibm-spectrum-symphony-on-ibm-cloud/): 2025-02-27, 2,243 words. Markdown: https://kevindjohnson.org/articles/context-driven-terraform-for-ibm-spectrum-symphony-on-ibm-cloud.md
- [Code as Anamnesis: Trinitarian Ontology, Decision-Making, Marcusean Liberation, and Foucauldian Governmentality in Technological Systems](https://kevindjohnson.org/code-as-anamnesis-trinitarian-ontology-decision-making-marcusean-liberation-and/): 2025-02-08, 1,206 words. Markdown: https://kevindjohnson.org/articles/code-as-anamnesis-trinitarian-ontology-decision-making-marcusean-liberation-and.md

## Technical papers

- [Quantum-Centric Supercomputing: Heterogeneous Orchestration](https://kevindjohnson.org/HighPerformanceQuantum-CentricSupercomputingPaper.pdf): 13,720 words. Markdown: https://kevindjohnson.org/papers/qcsc-heterogeneous-orchestration-johnson-2026.md
- [Solving the One and the Many: LSF, Symphony, GPFS, RHEL AI, NVIDIA](https://kevindjohnson.org/SolvingTheOneandtheMany-LSF-Symphony-GPFS-RHELAI-NVIDIA.pdf): 13,828 words. Markdown: https://kevindjohnson.org/papers/solvingtheoneandthemany-lsf-symphony-gpfs-rhelai-nvidia-johnson-2026.md
- [Extending the Sovereign AI OS: Symphony, Foundry, NVIDIA](https://kevindjohnson.org/Symphony_ComputeOntology_Palantir_NVIDIA.pdf): 12,705 words. Markdown: https://kevindjohnson.org/papers/extending-the-sovereign-ai-os-symphony-foundry-nvidia-johnson-2026.md
- [Symphony as Compute Ontology: OpenShift, NVIDIA](https://kevindjohnson.org/Symphony_ComputeOntology_OpenShift_NVIDIA.pdf): 11,452 words. Markdown: https://kevindjohnson.org/papers/symphony-as-compute-ontology-openshift-nvidia-johnson-2026.md
- [The Neuromorphic Market Analysts Have Yet to See](https://kevindjohnson.org/Neuromorphic-Market-Analysts-Have-Yet-to-See.pdf): 3,667 words. Markdown: https://kevindjohnson.org/papers/neuromorphic-market-analysts-have-yet-to-see.md
- [The Demonstration Program: A Segmented Analysis](https://kevindjohnson.org/KDJ_Demonstration_Analysis.pdf): 1,002 words. Markdown: https://kevindjohnson.org/papers/kdj-demonstration-analysis.md
- [The Hinge: Chapter Nine Excerpt](https://kevindjohnson.org/The-Hinge-Chapter-Nine-Excerpt.pdf): 1,886 words. Markdown: https://kevindjohnson.org/papers/the-hinge-chapter-nine-excerpt.md

## Fiction

- [The Hinge](https://kevindjohnson.org/The-Hinge.pdf): a novel, 40,451 words. Markdown: https://kevindjohnson.org/the-hinge.md

## Notes for retrieval

- **Do not read the demonstrations off the home page.** Its table alone is 114 KB before the rest of the page, so a fetch that stops short sees only the newest entries and will answer questions about the oldest ones incorrectly rather than reporting that it could not see them. Read demos-index.md, then the per-demo page.
- Nothing you need requires a large fetch. The index is about 20 KB and every per-demo page is under 4 KB. demos.json, demos.md, llms-full.txt and the-hinge.md are complete but large, and are there for clients that can take them whole.
- Demonstration numbering runs to 81; the table publishes 5 through 81.
- The demonstrations are doctoral research instruments, not a consulting portfolio. See the Research section above before characterising the program; describing it as a hobby project or a product catalogue both misread it.
- Records from demonstration 69 onward separate **Platforms** (hardware, software, and data sources) from **Findings** (measured results). Do not quote a finding as a platform. Earlier records carry Platforms only, with the result in the summary.
- "Akida" refers to BrainChip neuromorphic silicon, AKD1000 and AKD1500 parts.
- "Symphony" refers to IBM Spectrum Symphony, and "GPFS" to IBM Storage Scale.