Explore CUBE
Everything CUBE asks of you, before you commit.
Three rounds across four weeks. You build an AI agent alone, then make it survive inside a system built by four people you have never met, then defend the engineering in front of people who do this for a living.
The name
Four letters. One engineering culture.
CUBE is an acronym before it is a shape. Each letter is a thing we assess.
-
C
Collaborate
Work beyond your own code.
Understand ownership, dependencies, handoffs, interfaces, and how your work connects to everything around it.
Ownership · Coordination · Communication · Systems Thinking
-
U
Uplift
Make knowledge move.
Share what you know. Learn what you don’t. Teach, mentor, document, and grow through the people around you.
Knowledge Sharing · Mentorship · Learning · Leadership
-
B
Build
Demos end. Systems don’t.
Engineer autonomous systems that run, integrate, handle failure, evolve, and can actually be maintained.
Execution · Reliability · Architecture · Production Thinking
-
E
Earn
Proof over promises.
Build a record of demonstrated engineering, not a participation badge. Earn Sydon recognition through the work you actually deliver.
Credibility · Recognition · Achievement · Career Signal
The programme
Three rounds, then the defence
-
01 Apply by 16 September
Round 1: Selection
No quiz, no puzzle round. Applications are read for evidence of what you have actually built: your repositories, your commits, projects you can explain end to end.
-
02 25–26 September
Round 2: Build
One participant, one agent, one domain. You work inside a starter repository that already runs, so the first hour goes into the problem rather than the scaffolding. Submissions close 27 September: working code in the repository, merged.
-
03 7–9 October
Round 3: Integrate
The 100 finalists become 20 pods of five, one specialist per domain. Every agent must consume something another agent produces. The pod ships end-to-end workflows no single agent could complete alone.
-
04 10 October
Grand Finale: Defend
No slideware. The system runs in front of the judges, and then you defend it: architecture, decisions, failures, trade-offs, performance, and exactly what you contributed. Judges assess the system and the individual.
The five domains
Pick one. Own it completely.
Each domain is a real problem an intelligent commerce operation solves every day. You take one through Round 2, and in Round 3 it becomes one fifth of a working system. Selection opens 24 September, first come, first served.
Ocfqnpw Atgki
Ocf Qnp Watgkilbju
Ocfqnpwat Gkilbj
Ocfqn pw Atgkilbj
Ocfqnpw Atgkilbj
Key dates
Every date that matters
Friday 18 September, 9:00 AM through Saturday 10 October, 6:00 PM IST · Hyderabad, Telangana + online.
- 16 September 2026 Registration deadlineApplications close 11:59 PM IST.
- 18 September 2026 Kickoff9:00 AM – 1:00 PM IST · Hyderabad + online/hybrid.
- 24 September 2026 Challenge & problem statement releaseAll five domains open for selection, first-come, first-served.
- 25–26 September 2026 Round 2: build sprintIndividual build. One participant, one agent.
- 27 September 2026 Round 2 submission deadlineWorking code in the repository. Merged, or it didn’t happen.
- 28 September – 5 October 2026 Round 2 evaluationAssessed within each domain, so every pod gets one specialist per challenge.
- 6 October 2026 Round 3 finalists announced100 finalists · 20 pods · 5 specialists per pod.
- 7–9 October 2026 Round 3: integrationFive agents become one system with real cross-agent dependencies.
- 10 October 2026 Grand Finale: the defenceArchitecture, decisions, failures, trade-offs, performance and individual contribution.
Before you apply
The rules, stated plainly
- Who can enter
- Students, software developers and engineers, AI and GenAI enthusiasts, aspiring AI engineers, full-stack and backend developers, product engineering enthusiasts, AI-agent builders, and developers in startups and emerging tech.
- Do I need a team?
- No. Round 2 is entirely individual. Pods are formed for you in Round 3, from the 100 finalists, one specialist per domain.
- Can I use AI coding tools?
- Yes. ChatGPT, Claude, Gemini, Copilot, Cursor or equivalents. You remain responsible for understanding, testing, validating, explaining and defending everything you submit. That responsibility is the actual assessment.
- Online or in person?
- Both. The kickoff on 18 September is planned in Hyderabad with online / hybrid access; the build rounds run online. Venue details are confirmed closer to the date.
- What does it cost?
- Nothing. CUBE is free to enter.
- Is there a hiring outcome?
- Exceptional participants continue the conversation with the Sydon.AI engineering team. Hiring is an outcome of demonstrated ability across the whole programme, not a guarantee, and not decided by the final demo alone.
The operating system for modern commerce.