Founder Director & Sovereign Systems Engineer @ Kopano Labs · Cape Town 🇿🇦
Building evidence-first, offline-first systems that survive contact with reality.
✝️ Jesus is King. · POC before narrative · receipts before claims · governance before drift
My work sits where AI orchestration, adaptive web systems, persistent state, robotics, governance and South African infrastructure constraints meet.
The governing rule is simple:
If a system cannot show what changed, why it changed, what evidence supports the change, and what remains unknown, it is not ready to be treated as reality.
The August 2026 estate target is:
WORKING → CONNECTED → CURRENT → VISIBLE → EVIDENCED → BACKABLE
The semantic state of each system is governed in governance/profile-telemetry.json. GitHub activity is observational only. A scheduled GitHub Action regenerates the public SVG without allowing activity metrics to silently rewrite system truth.
OPEN // How the telemetry works
GOVERNED SYSTEM STATE
+
PUBLIC GITHUB ACTIVITY
↓
GENERATED SVG
↓
PROFILE SURFACE
stateis declared intentionally.pushed_atand stars are fetched as observations.- missing API data degrades to
activity unavailableinstead of inventing a value. - only
assets/estate-telemetry.svgis auto-committed by the workflow.
| System | What it proves | State |
|---|---|---|
| Kopano Context / Introduction-to-MCP | MCP-driven orchestration, KPGS governance, context authority, receipts and multi-agent workflows. | ACTIVE CORE |
| KasiLink | Township-first work discovery shaped by proximity, trust, utilities and low-friction access. | LIVE |
| Five's Arena · source | Production venue, bookings, fixtures, customer flows and live arena state. | PRODUCTION |
| Project Jennifer | A tactical RPG + governance simulator where relationship and world state persist through Memory Receipts. | R&D → VERTICAL SLICES |
| Cars4Mars | Rover safety/control, simulation, telemetry, edge perception and physical validation gates. | DESIGN → BUILD |
| AMAPHU | A governed entertainment surface connecting game, manga/anime, music and interactive media. | PHASED BUILD |
| Partial Knowable Algebra (PKA) | Private bounded-uncertainty runtime: X + Y = MAYBE, POCvsFOC classification and recovery loops. |
PRIVATE R&D |
Project Jennifer is one of the clearest visual and technical expressions of what I mean by governed intelligence:
choice → consequence → receipt → persistence → later gameplay
Public-lane Project Jennifer / Project Waifu Forge source assets. Visual existence is not treated as runtime canon without a separate governance receipt.
OPEN // Why Jennifer is more than an AI character
Project Jennifer treats continuity itself as gameplay. Important state should survive scene boundaries, conversation boundaries and restarts. AI can propose or interpret, but authoritative change is expected to resolve through governed state and receipts rather than silent narrative drift.
PLAYER CHOICE
↓
RELATIONSHIP / WORLD TRANSITION
↓
VALIDATION
↓
MEMORY RECEIPT
↓
PERSISTED STATE
↓
LATER SCENE REMEMBERS
I keep finding my handle in public datasets I did not author. Those records are useful as third-party discovery receipts, but their ranks and generated lists are mutable, so I do not treat a number as a permanent identity claim.
- 🇿🇦 gayanvoice/top-github-users — South Africa / total contributions — the generated table currently includes
RobynAwesome; positions can change as the dataset regenerates. - 🌍 xiv3r/top-github-users-ranking — an independent downstream ranking dataset also contains the account.
- 🤖 ishandutta2007/Top-AI-repos —
RobynAwesomeappears in the repository'sTBIindex. - 🧠 tejas-ae/Top-AI-repos — the README's public account/avatar index includes
RobynAwesome.
OPEN // Why I care about this
Self-description is useful, but externally generated public records give a different kind of signal:
BUILD → COMMIT → PUBLIC GRAPH → THIRD-PARTY INDEX → DISCOVERY
That does not prove that every ranking methodology is correct. It proves that the work leaves enough public surface area to be independently observed.
- Edge + experience: Next.js, React, TypeScript 7 target, Tailwind, Adaptive PWA, Three.js.
- Runtime: Node.js, Python, .NET 10 / ASP.NET Core, Turborepo.
- State: PostgreSQL, MongoDB Atlas, append-only receipts, offline-first local state.
- AI + orchestration: Model Context Protocol, Kopano Context, OpenAI / ChatGPT, Codex, Gemini, Copilot, Hugging Face, Ollama, Azure AI Foundry.
- Governance: KPGS, PKA, POCvsFOC, CI/CD evidence gates, provenance and recovery loops.
flowchart LR
A[Prompt / Human Intent] --> B[Proposal]
B --> C{Governance Decision}
C -->|approve| D[Artifact]
C -->|reject / unknown| H[Revise or Hold]
D --> E[Commit]
E --> F[Test + Measurement]
F --> G{Evidence Gate}
G -->|pass| I[Receipt + Visible State]
G -->|fail| H
I --> J[Next Context]
J --> A
OPEN // KPGS doctrine terminal
POC BEFORE NARRATIVE
RECEIPTS BEFORE CLAIMS
SOURCE AUTHORITY BEFORE RETRIEVAL CONFIDENCE
GOVERNANCE BEFORE DRIFT
UNKNOWN IS A VALID STATE
FAILURE MUST LEAVE A RECEIPT
PUBLIC VISUALS MUST NOT SILENTLY BECOME CANON
The point is not bureaucracy. The point is to keep powerful systems inspectable, recoverable and honest about what they know.
Currently: CPUT Computer Engineering Student · Johns Hopkins University Student · Robert Kennedy College
Building governed systems that turn concepts into receipts — from South Africa outward.
Profile evolved: 24 August 2026 · visual-first · evidence-first · living telemetry