Zain Dana Harper

Systems engineer & architect — AI accountability · compilers · real-time graphics · security · design

Seattle · independent since 2023/ harperz9.github.io/ github.com/HarperZ9/ zaindharper@gmail.com

Summary

Self-taught systems engineer with unusual range — compilers, real-time graphics, color science, AI accountability, and offensive/defensive security — unified by one discipline: work that proves itself. No claim ships without a test, a witnessed byte, or an inspectable artifact beside it; maturity is labeled, never inflated. The range isn’t scatter — it’s one conviction worked out in security and in creation at once. I build the accountability layer the agentic era needs, and I have the adversarial depth to harden it. Underneath the breadth is one operation I designed and named — the reconcile: perceive any artifact, check it against a criterion it did not author, carry a re-runnable proof, say UNVERIFIABLE when you can’t. Security and creativity are that same operation with a different criterion — which is why one architecture serves every domain. I architected and shipped all of it solo, directing AI as a coordinated engineering teammate — the systems thinking, the integration, and every verified claim are mine; the orchestration is how a body of work this wide ships from one person.

Selected work

Integrative architecture & delivery — the rare part

One abstraction proved by instances; a multi-repo program; AI directed as a coordinated teammate; and the design to present it all.

  • Abstract systems design — unified the whole stack under one operation (the reconcile) and demonstrated it across 15+ organs: the same loop is accountability’s spine, security’s gate, provenance’s witness, and creativity’s novelty-criterion. The kind of cross-domain abstraction rarely arrived at — and harder still to ship.
  • Program management — coordinated it as a multi-repo superproject under a spec → plan → build discipline: an ecosystem map and registry across ~100 logical projects, release engineering across many packages, every claim tracked to its receipt.
  • AI as a coordinated engineering teammate — architected and built solo by directing multi-agent systems (parallel agents for discovery, review, and construction) at a scale and breadth one person could not otherwise reach — while keeping the architecture, the integration, and every verified claim first-party. Directing AI at this level is itself a discipline.
  • Design & front-end — designed and hand-built the portfolio itself: a zero-dependency design system (semantic HTML / CSS / JS, WebGL, a live in-browser generative engine), proof-carrying and accessible — the presentation is a worked artifact, not a template. harperz9.github.io
The accountability spine — design + implementation

A composable stack built around one loop — perceive → gate → act → verify → witness — and the theory it implements.

  • Designed EMET, an external byte/provenance witness that answers MATCH / DRIFT / UNVERIFIABLE — never trusted; proved 19/19 conformance across three independent implementations (Python · Rust · Node).
  • Shipped proof-surface, a stdlib-only default-deny authorization gate with expiring least-privilege grants and delegation chains rooted in a real human; 258 tests, on PyPI.
  • Shipped coherence-membrane, zero-dependency perception organs (visual · raw · region · structured · audio · caption) that emit witnessed, re-derivable observations; 868 tests, on PyPI.
  • Built accountable-surface, the live loop — an MCP server with filesystem/web/command effectors, inert until authorized; 201 tests including a 39-test adversarial suite that forges digests, manufactures grants, and escapes bounds, confirming each is refused.
QuantaLang / quantac — typed-effects compiler
  • Designed a language where the effects a function may perform and the lifetimes of the references it returns are part of the type and checked through to native code; lowered end-to-end through a C backend; 1002 passed / 11 ignored.
Systems, graphics & color
  • Authored header-only C++23 signal/anomaly kernels — entropy, mutual information, Granger causality, PELT changepoint, FFT.
  • Built RAW (Rendering Advancement Workshop) — a public real-time D3D11 rendering platform: a proxy DLL that owns the pipeline and dispatches screen-space GI, AO, and reflections (GTAO · SSGI · SSR) mid-frame. Deep reverse-engineering + graphics; maturing toward the substrate's spatial engine.
  • Shipped color science on PyPI — 15+ spaces, CIECAM02/CAM16, perceptual ΔE, ICC, HDR tone mapping (quanta-color v1.0.1) plus a Windows display-calibration tool.
Security — build-and-break, for defense
  • Built a C++23 integrity / anti-cheat framework (~3,000 tests across ~59 modules) detecting tampering, evasion, and manipulation by understanding exactly how they are done.
  • Practiced binary reverse-engineering, instrumentation, and red-team research across the stack — repurposed inward to make the accountability platform self-accountable by adversarial construction.

Skills

Languages
PythonRustC++23HLSLWebGL / GLSLTypeScript
Domains
AI safety & classifier research · offensive security & AI red-teaming (jailbreak · social engineering · C2 / red-team tooling) · reverse engineering & binary analysis · real-time D3D11 / HLSL rendering (temporal AA · denoising · physically-based optics & bloom) · film color science & LUT pipelines · native driver dev (DDC/CI) · compiler internals & type systems · AI accountability & cryptographic provenance · agent orchestration · signal / information theory
AI dev tools
Claude CodeMCP server authoringagent orchestrationeval / red-team harnesses
Design & front-end
Hand-built design systems · semantic HTML / CSS · WebGL / Canvas · generative & algorithmic visual work · responsive, accessible, zero-dependency UI · typographic & visual design — the portfolio is the proof
Architecture & delivery
Cross-domain systems abstraction · multi-repo superproject coordination · spec → plan → build · multi-agent orchestration (parallel discovery · review · build) · proof-first release engineering
Practice
Disciplined release engineering — 6 packages on PyPI, more public on GitHub, each tested and released · test-first · proof-first documentation that labels its own limits

Selected receipts

19/19 EMET conformance 258 proof-surface 868 coherence-membrane 201 accountable-surface (39 adversarial) 1002 QuantaLang (11 ignored) ~3,000 integrity framework 6 on PyPI (+ GitHub)

No computer-science degree — independent since 2023, out of Seattle. What’s here was architected and shipped solo, in the open, in a remarkably short span — the breadth, the integration, and the rate are the commitment. The work is the credential, and it is all inspectable: harperz9.github.io. · Updated 2026-06-23.