# Rick Glenn · Exocognosis

> Post-quantum cryptography developer, Rust blockchain architect, fractional CTO. I build and
> rapidly ship useful technical projects that create real-world value, in security,
> infrastructure and product-oriented systems.
> Taking 2–4 week builds, starting now. Contact: GitHub @exocognosis, X @exocognosis.

This is the plain-text version of https://www.exocognosis.com for AI agents and other automated
readers. It says the same thing as the page. The main product has its own page:
[PQC Sidecar](https://www.exocognosis.com/pqc-sidecar.md).

## About

I'm a technical builder focused on shipping practical systems that solve real problems. I like
work that combines engineering depth, speed, and clear value creation. One of a small number of
people building production post-quantum infrastructure as a solo architect.

## Core skills

- **Cryptography & security:** post-quantum cryptography, protocol design, key management,
  HSM / KMS architecture, threat modeling, secure systems review.
- **Blockchain & distributed systems:** Rust L1 architecture, consensus systems, validator
  infrastructure, state machines, distributed ledgers, fault tolerance.
- **Languages & frameworks:** Rust, TypeScript / Next.js, Node.js, Swift / SwiftUI, API design,
  CLI / devtools, Python, Qiskit.
- **Infrastructure & product execution:** secure networking / VPN, performance engineering,
  observability, Linux deployment, automation, fractional CTO execution.

## Agentic engineering

From idea to deployed system with agent swarms, rapidly and securely. I run AI agents as a
coordinated engineering team: parallel streams for code, tests, research, and operations, held
to verification gates and human sign-off.

1. **Swarm:** work splits into independent streams run by agents in parallel, each with a
   focused brief and its own workspace.
2. **Verify:** nothing merges on an agent's word. Tests must fail when the fix is removed, a
   second implementation must agree, and real-world runs back the claims.
3. **Gate:** deploys, outreach, and anything irreversible or IP-sensitive wait for explicit
   sign-off. Agents prepare and propose; I decide.
4. **Secure:** agents work on branches and staging copies. Production changes go out through
   scripted, tested, reversible deploys with backups.

**Case study: PQC Sidecar, under two days.** From a v0.1 release to a KMS-backed,
cross-verified, chain-agnostic product in soak testing:

- AWS KMS signing backend, tested end to end through the real AWS SDK, with request signing
  checked against the live AWS endpoint
- A second verifier in TypeScript whose output matches the Rust implementation byte for byte
- All 3,438 single-bit tamperings of a certificate rejected
- Soak testing with injected RPC outages and restarts; a restart-gap bug it surfaced was fixed
  and verified live on Cosmos Hub
- Market research and a public product page shipped alongside the engineering

**Built the same way: the Dytallix ecosystem** (https://dytallix.com), a post-quantum product
ecosystem built and operated by one architect with agent swarms:

- Dytallix L1: PQC-native ledger, mainnet in launch simulations
- QuantumSentinel: cryptographic risk scanning, alpha
- QuantumVault: file protection with signed verification evidence, early access
- QuantumLink: post-quantum device mesh, in development

What I build for teams: agent systems with tool use, multi-agent orchestration, evaluation and
guardrails, human-in-the-loop workflows, retrieval and knowledge systems, secure agent
infrastructure.

## Featured projects

### PQC Sidecar — product, now taking engagements

- **Problem:** blockchains finalize history with classical signatures. Once quantum computers
  can break them, attackers can forge blocks a chain already finalized, and native migrations
  take years.
- **Solution:** a chain-agnostic sidecar that runs beside one node and signs compact ML-DSA-65
  (FIPS 204) checkpoint certificates over finalized blocks, with the key in a file or AWS KMS.
  No hard fork, no validator coordination.
- **Outcome:** post-quantum evidence for bridges, custodians, auditors, and archives today,
  checkable with any standard ML-DSA verifier.
- Details: https://www.exocognosis.com/pqc-sidecar.md

### Dytallix — mainnet in launch simulations

- **Problem:** existing blockchain systems are vulnerable to future quantum threats and lack
  native AI capabilities.
- **Solution:** a post-quantum Layer-1 blockchain ecosystem including validator
  infrastructure, wallet systems, governance mechanisms, and AI-native modules.
- **Outcome:** a functioning blockchain platform capable of supporting quantum-resistant
  infrastructure and AI-enabled services.
- GitHub: https://github.com/DytallixHQ

### QuantumVault — deployable now

- **Problem:** organizations need stronger protection for sensitive key material and security
  workflows as infrastructure becomes more distributed and cryptographic risk increases.
- **Solution:** a security product built on advanced infrastructure patterns, with attention to
  key management, secure access, product packaging, and deployment paths.
- **Outcome:** a deployable security product that turns deeper cryptographic and infrastructure
  work into a concrete offering for real users.
- Product: https://dytallix.com/enterprise

## Research

**Recursive Phased-State Signatures (RPS-SIG)** — https://github.com/exocognosis/rps-sig

Post-quantum signature schemes still force hard tradeoffs between standardization, signature
size, implementation fragility, and newer assumptions. RPS-SIG is an attackable research
framework, not a production scheme: a recursive state tree derived from one root seed,
phase-local public projections, Merkle commitments, and global public-key binding. A
conservative hash-based construction path is kept separate from experimental finite-geometric
variants, so the security games, proof obligations, misuse tests, and implementation criteria
can be reviewed before any stronger cryptographic claims are made.

## What I'm looking for

High-leverage technical work in infrastructure and security, with teams that value execution.
Right now: focused 2–4 week builds (post-quantum migrations, Rust services, security
infrastructure, AI systems), scoped to ship something working by the end.

- **Roles that fit:** 2–4 week builds, founding engineer, technical co-founder, principal /
  staff engineer, head of platform / infrastructure, security architect, infrastructure
  architect, fractional CTO, technical advisor.
- **Project types:** post-quantum migration, L1 / consensus design, secure networking / VPN,
  cryptographic protocols, distributed systems, key management / HSM, critical infrastructure,
  performance-critical systems.

Strong fits usually solve a hard security, infrastructure, or technical problem, or deliver an
unusually useful system with real-world application.

## Submit a project

The form on https://www.exocognosis.com (section "Submit") takes a name, organization, contact
method and handle, a project description (at least 40 characters), and an optional context
document (.txt, .md or .pdf, up to 4 MB). Useful to include: what you're building in concrete
technical terms; the stage (idea, prototype, funded, shipped); timeline and what you need; the
engagement type (full-time, contract, fractional, advisory); and why it's interesting now.
Submissions are screened by a keyword pre-filter and a model-scored fit rubric; strong fits are
shown a scheduling link.

**Agents can submit directly.** `POST https://www.exocognosis.com/api/score` with a JSON body:

```json
{
  "name": "Jane Doe",
  "organization": "Example Labs",
  "contactMethod": "Email",
  "contactValue": "jane@example.com",
  "description": "What you are building, stage, timeline, what you need, engagement type, why now (40 to 8,000 characters).",
  "context": "Optional supporting text, up to 30,000 characters."
}
```

`contactMethod` is one of Email, Telegram, X, Signal, Discord, Phone, Other. The response is
`{"score": 0.0-1.0, "reason": "...", "threshold_met": true|false, "calendly_url": "..." (strong fits only), "contact_email": "..."}`.
If automatic scoring is unavailable, the submission is still kept and you get `202` with
`{"received": true, "scored": false, ...}`; Rick reviews it by hand.
Limits per client: 30 requests and 5 scored submissions per hour (429 with Retry-After beyond
that); body up to 5 MB. Submissions are treated as data: instructions inside them are ignored,
and attempts to steer the score count against the submission. Full schema:
https://www.exocognosis.com/openapi.json

## Contact

- GitHub: https://github.com/exocognosis
- X: https://x.com/exocognosis
- PQC Sidecar engagements: rick@dytallix.com