Robert Lingoes · Founder, Arcturion Group
I began building with AI in February 2026, knowing very little about it, and everything I have learned since, I have taught myself. Arcturion is the company I have built in that time, run in partnership with a team of AI agents I designed myself. Where this technology is heading does not unsettle me. It inspires me.
Built in public, directed by me and coded with AI.
Selected work
Each began as a practical difficulty in my own working week.
Every new conversation with an AI once began from nothing, and I grew weary of re-explaining my world. The Arcturion Universe gives my agents a shared memory and clearly defined responsibilities, while every consequential decision returns to me.
Read the case study
For seven years I kept Konexion’s books myself, and throughout that time I wanted accounting software capable of explaining why a number is what it is. Arcturion Finance is that software, and I am building it now.
Read the case studyArcturion Technologies
Each is published on GitHub for anyone to adopt.
Whenever I solve a problem for my own agents, I ask whether the solution might serve someone else. If it can, I publish it on GitHub. I want more people building seriously with AI, and I am glad to put my name to that work in public.
Explore TechnologiesThey include a generator that keeps agent instructions synchronized across Claude Code, Codex, Gemini and other environments; hooks that keep agents accurate about the current time and proportionate in the length of their answers; connectors for Spotify and Apple Reminders; a router that assigns modest tasks to smaller models; and a meter that accounts for every token consumed.
My public repositories Arcturion Technologies on GitHubArcturion Capital
He currently trades simulated capital only.
Vega studies the research, tests every idea before it may approach a trade, and keeps a complete record, losses included. I also asked him to explain his reasoning in plain English, because the language of finance can obscure as much as it reveals.
Meet VegaVega trades simulated capital. This is education, not investment advice.
How I build
AI writes the code. My responsibility is to determine what deserves to be built, to articulate precisely what it must accomplish, and to review the result with an exacting eye. Where it falls short, I say so, and we refine it. Two decades spent understanding what clients truly want proved to be an unexpectedly rigorous preparation for directing artificial intelligence. I hold no formal training in technology; I am proudly self-taught, and nearly every free hour I have is given to reading, taking notes and training alongside my agents.
Meet the founder →If a project would not resolve a real difficulty in my own work, I do not build it.
It sets out what the agent owns, what it requires from me, and when it must return for direction.
That means unusual cases and malformed inputs, precisely the conditions a polished demonstration never reveals.
Collaboration
I am drawn to a workflow no one can follow, an archive of notes in which nothing can be found, a process that consumes an entire Monday, and the idea that will not leave you alone.
From me, you would receive a brief we have agreed upon together, a clear map of how the work actually flows, your knowledge organized so that AI can put it to use, and a candid review of everything the AI produces.
Discuss a collaborationGet in touch
I welcome conversations with founders, operators, teams and those taking their first steps with this technology. Few subjects engage me more.
Send me a message