FAROS is my research into multistrom satellite clusters. FAROS introduces the stroma in space: a distinct region of orbit, at its own altitude, with its own character and purpose. For example, picture a slice of bread spread with butter and marmalade. Each keeps its own character, yet together they make one whole; blend them into a paste and that character is lost. A multistrom network works the same way. One stroma relays traffic, another carries the backbone, another stores data, and all of them connect in space. It is not optimal to keep all types of satellites and functions in all orbits. I simulated it in Python, routed traffic using Dijkstra's algorithm, and built an interactive 3D Earth to test it.
Published in the National High School Journal of Science, October 2026.
Research first, then the hardware, then the software.
Surface data routing is fragile: submarine cables have been severed several at once, repairs take days, and routes are set by geography and landing rights. Orbit is subject to none of that.
The result is that building in orbit is requirement and not only optimization. On an intercontinental route, nothing that comes back down to Earth can close the distance. The network connects in orbit or it does not connect.
Speed is the conditional part. A backbone at 3,000 km cuts the route from five satellites to three, but the signal travels 4,090 km further to get there — a trade that pays above 6.80 ms of handling per satellite. That is a hardware number, and it moves toward the backbone as links shorten: 2.70 ms at a 2,000 km link range, and shorter still until the backbone is the only connected path left. Delay is also not one fixed number. The satellites keep moving and the Earth turns beneath them, so the best route is constantly being replaced — roughly every minute — and the time a signal takes drifts with it. A higher stroma uses fewer hops, which is a reason to expect steadier routes. More satellites is not the answer by itself: spread the same satellites across more rings around the Earth and the network improves — then saturates at about ten.
In my first version of FAROS, I kept only what the simulation could support, so I removed the political conclusions I had drawn. I still see, however, Spacepolitics as inevitable when decisions about space are made, and a future version of FAROS will be about it.
I am the 3D designer for the Kirineos data centre at MTEL. I model parts in SolidWorks, produce the manufacturing drawings, and send the CAD to factories, where the parts are made and installed.
My capstone for the MITxPRO Professional Certificate in Full Stack Development with MERN — 26 weeks, a deliverable every week — is a Next.js and Postgres dashboard with authentication, server-side validation and a settings module I built beyond the course material.
Before that: a robot guitar in Python at ten, an interactive mapping tool, a health-tracking dashboard, and a vocabulary trainer. These are where I learned to write code well enough to do the research.
See the software