HQDigital headquarters
This is my digital headquarters: the projects I ship, the AI agents I build, and the product ideas I'm working through. Everything here is either real and inspectable, or labelled as what it is.
01About
Developer · AI builder · Systems thinker · Entrepreneur
I do more than write code. I explore ideas, build intelligent systems, automate workflows and develop products. Most of what I make starts as a question (could this be automated, explained better, or turned into a product?) and ends as something you can open and use.
My public work so far includes Researchora, an AI-assisted research community. More is in the workshop and will appear here when it's ready to share.
If something works, you can try it. If it doesn't yet, it says so. Demos are labelled as demos and prototypes as prototypes.
Numbers and claims should lead back to a source. Assumptions are fine as long as they are marked as assumptions.
Models are for language and judgement. Anything that can be calculated or checked deterministically is done in code.
Agents can propose and prepare. Publishing, sending, spending and changing production systems wait for an explicit yes.
02Projects
Every repository here is fetched from my GitHub account, so stars, languages and dates are current. Download a ZIP, copy a clone command, or have the README explained first.
Researchora
A frontend-first research and knowledge-sharing platform: discovery, publishing, discussion, reading history, following and an AI assistant.
Not open source. All rights reserved; use beyond personal evaluation needs written permission from the owner.
ConnectedResearchora AssistantREADME ExplainerVenture: Researchora
gaze-companion
No description has been written for this repository yet.
No license file. The code is public to read, but without a license it isn’t free to reuse.
The_food_app
No description has been written for this repository yet.
No license file. The code is public to read, but without a license it isn’t free to reuse.
Live from the GitHub API · data as of Nov 5, 2026 · All repositories on GitHub
03AI Lab
Each agent has a profile: what it is for, what it can and cannot do, what it is built on and how a request moves through it. Some run on this site and can be tried below; others live inside my projects. Every one carries an honest status.
On-site AI runtime
Checking…
Answers questions about this portfolio: the projects, agents and ventures in it.
Recommends the work most relevant to what a visitor says they care about.
Reads a public repository's README and explains what the project is and how to run it.
Turns pasted free text into structured records that software can use.
Walks a visitor through the reasoning behind one of the ventures on this site.
Breaks a product idea into the assumptions it depends on and how to test them.
The AI assistant inside Researchora, for asking research-related questions.
Talk to the agents that run on this site. Replies stream from a real model call on the server.
HQ Guide
Answers questions about this portfolio: the projects, agents and ventures in it.
HQ Guide
Ask me about anything in this headquarters. I answer from the site's own project data, so if something isn't listed here I'll tell you rather than guess.
Try asking
Answers are generated from this site’s own data and can still be wrong. Conversations aren’t stored.
How I structure work across several specialised agents: narrow roles, explicit hand-offs, and a human gate before anything consequential. Select a component to inspect it, or step through the flow.
Five specialised agents take a question from open-ended research to a reviewed, approved action. It shows how I structure agent work: narrow roles, explicit hand-offs, and a human gate before anything consequential.
Example task: “Find out why setup is confusing for new users of a repository, and propose a fix.”
Step 1 of 6scripted walkthrough
Reads the README, the open issues and the setup scripts. Each finding is stored with the file or issue it came from.
Inspecting
Gathers relevant information
Guardrail. Retrieved content can inform the answer but can never change the agent's instructions.
A few focused tools rather than many shallow ones. Each states what it accepts, what it returns, what powers it and where it stops.
04Venture Studio
Where technical work meets a real problem. Each venture has a profile covering the problem, who has it, the proposed solution and what still needs to be proven.
Each venture sits at the stage I’ve set for it by hand. Nothing here is launched until it says “Launched”, and no traction is claimed unless it is real.
SaaS product
An AI-assisted research community: discover, publish, discuss and follow research in one place.
05Teaching
Building things is half of it. The other half is helping other people get started. In summer 2026 I founded and taught a live online program for middle and high school students.
A live online program I founded to teach Python programming and artificial intelligence to middle and high school students. I designed the curriculum and taught every session myself.
After the program I set up a parent media-permission process, so student projects could be showcased on Instagram, TikTok, YouTube, Facebook and LinkedIn while protecting students’ privacy.
06Progress
Small, true numbers. Each one names its source.
Public repositories
3
Source: GitHub API
GitHub stars
3
Across public repositories
Source: GitHub API
AI agents documented
7
6 run on this site
Source: maintained by hand
Ventures
1
1 completed
Source: maintained by hand
Launched products
0
Nothing has launched yet
Source: maintained by hand
Site visits
–
Counting…
Source: this site’s counter
Share of bytes across 3 public repositories
Source: GitHub API
The status I assert for each agent in the AI Lab
Source: maintained by hand
Dated events, newest first
GitHub figures are live as of Nov 5, 2026. Agent and venture figures count the entries published on this site. A visit is one browser session; bots are left out, and no cookies or IP addresses are stored. This dashboard shows no revenue, customer or funding numbers because none are tracked here.