A global network of AI hackathons
Two days. One working prototype. Every city, the same bar.
11 editions across 10 countries, all running the same two days, the same hundred-point card and the same two-round judging — so a score in one city means what it means in another.
Founded by Starnest Academy with the OMNI AI Summit. Free to take part, subject to selection.
- Editions
- 11
- Countries
- 10
- Places
- 2307
- Points
- 100
across 9 regions
on four continents
across the network
the same card everywhere
One standard, every city
Editions are run by the network, licensed to local hosts, or delivered white-label for a client. All three are judged on the identical card, in two rounds, by judges who cannot see each other's sheets.
- Applications open
- Live
- Announced
- Completed
- Announced
Tashkent
Uzbekistan · Central Asia
19–20 February 2027
- Announced
Istanbul
Türkiye · Europe / MENA
26–27 March 2027
- Announced
Dubai
United Arab Emirates · GCC
23–24 April 2027
One edition has run. The rest are announced dates, published early so hosts and partners can plan against a calendar — not confirmed bookings.
What makes it a network and not a franchise
Eight obligations every edition meets before it may use the name. These are the parts a local organiser is most tempted to drop, which is exactly why they are the ones written down.
- 01
Two days, in person
Day one is team formation and building, and ends with the pitch deck deadline. Day two is judged: the first round in the morning, the final and the awards in the afternoon. No remote-only editions — the format depends on the room.
- 02
Teams of one to five
You can enter alone or bring up to four other people. Teams form before or during day one, and five is the ceiling, not a target.
- 03
The same hundred points
Value for the user 25, prototype and use of AI 30, quality testing 20, feasibility 15, originality 10. Published before registration opens, identical in every city.
- 04
Two rounds, never one
Every project is read in a first round against the card, then the finalists are rescored from scratch by a human jury. How the first round is read is the edition’s choice — Baku has GPT and Claude do it live on stage. A single round cannot be compared across cities.
The card, everywhere
- Value for the user25
- Prototype and use of AI30
- Quality testing20
- Feasibility15
- Originality10
Thirty-five points are not the demo
Quality testing and feasibility together outweigh the prototype. A two-day AI demo is easy to make impressive and hard to make honest — the weight sits in that gap, in every city.
Ties resolve by the prototype score, then quality testing. Judges declare conflicts and are replaced rather than quietly excused.
Two tracks, one final
Both tracks run on the same hundred points, and the finalists are taken across the two of them.
AI Gaming
Artificial intelligence applied to games and the games industry. Change how a game is played, made or operated, and show the scenario running — a trailer and a slide deck are not a prototype.
AI Enterprise Solutions
AI products for business: automating a process and solving a corporate problem. Take one workflow a company runs today, show your prototype doing it end to end, and name the metric that says it is worth adopting.
Who is behind it
- AZ
Azercell
Supporting partner. Proposed scope: specialists as mentors and judges, and support with prizes, tooling and connectivity.
- OM
OMNI AI Summit
The stage where the winners are announced on 10 October.
- SA
Starnest Academy
Organiser. Runs the programme, participant selection, mentors and judges, submissions and results.
It starts in Baku, 9 October
The founding edition runs at the Baku Convention Center. Everything the network does afterwards is measured against what happens in that room.