ABOUT THE CHALLENGE

Design and deploy self-improving AI agents that can optimize their own behavior over time. Join the region’s top AI builders, researchers, and engineers for an intense, in-person competition focused on the future of autonomous systems.

We Bring:

  • ​Expert judges
  • ​Good food
  • A great set-up for your team with our partners at AUD

​You Bring:

  • ​Your Ideas
  • ​Positive Vibes
  • A readiness to crank out code (your own laptop)

What's the project you've been sitting on? A research agent? A library for the agentic era? An agent training framework? Whatever the idea, it's time to stop talking and start coding.

If you are a student, creator, researcher, or extreme enthusiast in AI, autonomous agents, self-optimizing systems, multi-agent workflows, and everything in between... this is your green light!

Register to build.

 

​What Will You Build?

​Participants will design self-improving agent systems that can: 

  • ​Learn from data and feedback
  • Adapt strategies over time
  • Optimize workflows, decisions, or behaviors autonomously

You may integrate any advanced tooling such as agent platforms, voice APIs, memory systems, evaluation layers, or workflow orchestration tools etc., of your choice or provided by partners.

 

EXPECT:

  • ​Win up to 100,000 AED in cash and partner prizes
  • One-day, in-person challenge
  • Work on real-world agent systems
  • ​Learn from expert mentors in AI and autonomous systems
  • Collaborate with elite engineers, researchers, and startups

SCHEDULE:

 

​9:00 – 9:30 | Registration & Check-In

9:30 – 9:45 | Opening Ceremony & Partner Talks

10:15 – 10:30 | Challenge Briefing + Judging Criteria + Q&A

10:30 – 4:45 | Hacking Session

12:30 | Lunch Break

4:45 | Code Freeze

5:00 – 6:00 | Pitches

6:00 – 6:45 | Closing + Partner Keynote + Awards

 

PRIZE POOL:

27,500 USD prize pool.

Sponsored by Ruya AI and our partners ClickHouse Amazon Web Services (AWS) Datamellon ElevenLabs.

All participants receive access to AWS Infrastructure on the 14th and one month of ElevenLabs Creator tier. 


SPONSORS AND PARTNERS:


US Consulate - The U.S. Consulate in Dubai represents U.S. government interests, supports American citizens, and promotes economic, educational, and cultural cooperation between the U.S. and the UAE.

AWS - A global leader in cloud computing, providing the secure and scalable infrastructure that powers today’s most innovative businesses. We are proud to have them as our sponsor, helping us support the next generation of creators and engineers.

American University of Dubai - Leading private university in the UAE, known for American-accredited programs and strong links to industry, innovation, and entrepreneurship.

ElevenLabs Hacker Guide

https://docs.google.com/document/d/1mCh5MtOzBw0aJpurQVUmIVFfPMAHW3MjNemE-LNiMto/edit?usp=sharing

 

 

Applying To Participate - Who Should Apply?

  • ​Top AI & Engineering Students from leading universities.
  • ​Early-career engineers and researchers at AI startups and labs.
  • ​Developers building agent systems, orchestration tools, or applied AI.

​We use an application-based selection process to ensure a high level of technical depth and meaningful collaboration.

Join Us - APPLY NOW!

Requirements

The Challenge: Self-Improving AI Agents

 

  1. Register and create a self improving AI agent in any industry, vertical or space.
  2. The idea and code must be yours. Give credit where credit is due for APIs, assets, etc.
  3. All code and meterials must be produced within the hackathon time.
  4. Vibing with Claude Code, Codex, Cursor or any model/agent/ide is allowed.

Submissions:
  1. Link to a public GitHub repository of the project. [Refer to My Projects -> <Your_Project_Name> -> Project Details -> "Try it out" links]
  2. Provide at least ONE of the following:
    - A link to a YouTube demo video (maximum 3 minutes). [Refer to My Projects -> <Your_Project_Name> -> Project Details -> Video Demo link]
    - A PDF presentation outlining your deliverables (suitable for a 3-minute presentation). [Refer to My Projects -> <Your_Project_Name> -> Additional Info -> Upload a file 35 MB max.]

Hackathon Sponsors

Prizes

$26,500+ in prizes
+ other prizes
1st Place
$11,500 in cash
1 winner

10,000 USD in Cash.

Sponsored by Ruya AI and our partners ClickHouse Amazon Web Services (AWS) Datamellon ElevenLabs.

1,000 USD in AWS Credits.

1,800 USD in ElevenLabs Pro - 6 months Pro subscription for each team member.

2nd Place
$8,500 in cash
1 winner

6,000 USD in Cash.

Sponsored by Ruya AI and our partners ClickHouse Amazon Web Services (AWS) Datamellon ElevenLabs.

1,000 USD in AWS Credits

1,800 USD in ElevenLabs Pro - 6 months Pro subscription for each team member.

3rd Place
$6,500 in cash
1 winner

4,000 USD in Cash.

Sponsored by Ruya AI and our partners ClickHouse Amazon Web Services (AWS) Datamellon ElevenLabs.

1,000 USD in AWS Credits

1,800 USD in ElevenLabs Pro - 6 months Pro subscription for each team member.

Antler Dubai Residency
1 winner

Antler will award one standout team a direct seat in their April 6th Dubai Residency, giving each participant the chance to build with Antler from day zero, independently or together, with access to up to $500K in potential investment.

10,000 USD in AWS Credits
1 winner

10,000 USD in AWS Credits - Amazon Web Services will support a selected team not just at the prototype stage, but all the way to building a real company.

As they take the next step and incorporate in theUAE, AWS will award $10,000 in credits to help accelerate development and scale what they've started.

Devpost Achievements

Submitting to this hackathon could earn you:

Judges

Arslan Berbic

Arslan Berbic
Ruya AI

Junaid Qureshi

Junaid Qureshi
Ruya AI

Sako M
Independent

Akash M
Independent

Ezzat Mohamed

Ezzat Mohamed
AWS

Tyler Hannan
ClickHouse

Michael Jacobs
Independent

Siddhant Dayal

Siddhant Dayal
Palantir

Judging Criteria

  • Completion and submission on time
    Participant shall stay on the topic, and complete the aforesaid assignment within the allocated timestamp
  • Building Self Improving Agents
    Judging: 3-min live demo per team (strict). Demo first, max 2 slides. Judges ask quick questions. Scored on self-improvement, agent design, real-world impact, easy setup/run, and demo clarity.

Questions? Email the hackathon manager

Tell your friends

Hackathon sponsors

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.