Monitor. Adapt. Defend.
Info
Are you a student or a professional passionate about AI and LLMs?
The 6-hour AI Agent Challenge is a team competition for Replyers passionate about LLMs and AI. Teams will design a multi-agent system for a complex fraud detection scenario, focusing on agent creation logic and building intelligent, dynamic systems. Domain expertise is not required.
00 days
00 hours
00 minutes
To play the challenge registration in a team is needed.
Register by 11th March, 23:59 CET.
By joining the challenge, you’ll get hands-on experience designing and testing AI agents in a realistic fraud detection scenario, working with systems that mirror real-world complexity in a safe, protected environment.
To compete you have to form a team from 2-4 members.
If you already know some good AI experts, create your own team and invite your friends.
Otherwise search for and meet new teammates from all over the world on the Discord server.
⚠️ If you don’t join any team, your registration won’t be completed and you cannot take part to the challenge.
There is always room for improvement. Access exclusive learning material in order to collect knowledge and hints to face the competition as a real code master.
On March 21, at 15:30 CET, we’ll publish the problem statements on your challenge page and you'll have 6 hours to play.
A leaderboard will be regularly updated to show how your and other teams are performing.
One week after the challenge day the final leader board will be published and all finalists for each category notified.
Each member of the winning team wins:
🥇 An iPhone 17 Pro
🥈 A PlayStation 5 Slim Digital Edition
🥉 A pair of RayBan Meta - Gen 2
The challenge has been designed by a team of 20 AI experts from Reply, that will support you on the day of the challenge.
First of all check your mailbox regularly: you’ll get some emails before and after the challenge. On this website you can monitor your challenge progress directly in your profile. If something won't be clear, you can always message the Reply AI-Vengers during the challenge via chat.
Every submission must include a file containing the source code of the program you developed to generate the output. Each source code file must be an ASCII plain text file or an archive of ASCII files. Each output file must be an ASCII plain text file. All input files will be an ASCII plain text.
At the end of the challenge, the Reply AI-Vengers will review and validate the best scoring submission from top-ranked teams on the leaderboard.
