ZeroBandwidth
zb@zerobandwidth:~$whoami

ZeroBandwidth

Systems Architect/Veteran/AI Builder/Cybersecurity Researcher/First-Principles Engineer

I build strange, useful, real systems at the edge of what most people think is possible — from local AI cognition and synthetic clinical reasoning engines to AI education labs, code-rescue networks, robotics perception, and adversarial security tools. This is the canonical home of ZeroBandwidth.

Canonical Identity · Verified

zerobandwidth.com is the official home of David's ZeroBandwidth identity.

Verify identity →
02Project Constellation

A map of the work, not a list of links.

Major systems grouped by domain — AI cognition, education, trust, founder tools, and experimental R&D. Hover the nodes. Walk into the lab.

ZB://constellation.map
03Featured Systems

The headline builds.

A cross-section of what lives here — from cognitive architecture to trustworthy agents to shipped SaaS.

COGNITIONResearch

JARVIS — Oracle Edition

Real-time perception + memory + bounded reasoning

A local AI cognition system built around perception, memory, world modeling, bounded autonomy, and epistemic defense.

Local AIRoboticsCognitive ArchitectureMemory Graph
View in constellation
EDUCATIONLive

hackme.ai

Real learning systems, not fake AI demos

An interactive AI education lab where people learn machine learning by breaking it, training it, and watching systems fail honestly.

AI EducationNeural NetsReinforcement Learning
View in constellation
TRUSTPrototype

MirrorIQ

Hold a system up to its own reflection

A trust and validation system that mirrors a target back to itself to expose inconsistency, drift, and adversarial weakness.

TrustValidationAdversarial AI
View in constellation
TRUSTLive

NjordScan

Security scanning that respects the developer

A security scanning tool focused on surfacing real, actionable issues without drowning the developer in noise.

SecurityScanningDevSec
View in constellation
TRUSTLive

Trustworthy AI Agent Blueprint

How to build an agent you can actually trust

A canonical, evolving blueprint for designing AI agents with bounded autonomy, validation, and defensible behavior.

AI AgentsTrustArchitecture
View in constellation
FOUNDERLive

SoloNerds

Infrastructure for the technical solo founder

A platform and toolset for solo technical founders shipping real products without a team behind them.

SaaSFounder ToolsCloud
View in constellation
R&DResearch

Clinical Reasoning Engine

Synthetic clinical reasoning, made auditable

A prototype platform for synthetic clinical reasoning — structured, explainable inference over medical decision-making.

Clinical AIReasoningExplainability
View in constellation
04Proof, not vibes

Real systems. Real architectures. Real failure modes.

Anyone can use AI. The hard part is layering systems that survive contact with reality — because you understand how each layer fails before you stack the next one on top.

35+Years building systemsHardware, networks, software, and the seams between them.
TS/SCIRetired military technologistCleared background. Systems where failure was not an option.
CTOExecutive technical leadershipOwned architecture and outcomes, not just code.
10+AI platforms & systems builtCognition, education, trust, and agent architectures.
LiveShipped products in productionReal users on real infrastructure, not slideware.
OpenOpen-source repositoriesCode in public, under the Nimdy handle.
05Field Notes

The public record.

Build logs, lessons learned, and the things I was right about early.

All field notes