About the Competition
Build an AI poker agent and compete for $6,000 in prizes. Test your skills in machine learning and game theory against other university students.
Career Opportunities
Direct recruiting opportunities with Jump Trading for top performers. Network with industry professionals and showcase your skills to potential employers.
Hybrid Format
Compete online during the main competition. Opening ceremony, closing ceremony, office hours, workshops, and recruiter events all held in-person on campus.
Prize Pool
Competition Rules
Eligibility
- Open to all CMU students
- Teams of up to 2 members allowed
Technical Requirements
- Agents must be implemented in Python
- Each bot runs with 1 vCPU and 2 GB of RAM
- Maximum runtime of 7 minutes per agent per match
- Use of external libraries limited to approved set
- Additional libraries may be requested for approval
Match Format
- Matches are heads-up (1v1)
- Each match consists of 1000 hands
- Players start each hand with a fresh bankroll
- Winner is determined by total bankroll across the match
Tournament Structure
- Teams play matches automatically throughout the day during the online phase
- Match requests can be made to higher elo opponents
- Teams with the highest elo win the competition
Competition Timeline
Registration Opens
Team registration begins. Start forming your teams and preparing your development environment.
Monte Carlo Workshop
Learn about Monte Carlo tree search and other poker algorithms to prepare for the competition.
Poker/Game Theory Workshop
Learn about poker and game theory to prepare for the competition.
Competition Begins
Access to the variant, competition framework, and documentation announced at the opening ceremony. Office hours begin.
Competition Ends
All teams must submit their final AI agents by 11:00 AM Mar. 29 EST.
Closing Ceremony
Winners are announced at the closing ceremony.
Frequently Asked Questions
What programming languages can I use? +
You can use Python 3.12. Many common external libraries will be available for use. If you'd like to use any others, you can request them in our Discord.
How many people can be on a team? +
Teams can be between 1 and 2 CMU students.
What poker variant will be used? +
The specific poker variant will be revealed when the competition starts. It will be more complex than standard no-limit Texas Hold 'em.
What are the prizes? +
The prize pool will be $3,000 for first place, $2,000 for second place, and $1,000 for third place.
Is there a registration fee? +
No, participation in the competition is completely free.
What frameworks are supported? +
You only need to implement a simple interface we provide that tells your bot when it's your turn to act. You're free to use any approach you want - from simple rules to deep learning. For those interested in reinforcement learning, the game engine also implements the OpenAI Gym interface, and we'll provide example code for getting started with RL approaches.
How does matchmaking work? +
Matches will be run continuously with approximately 20 matches per day per team. Teams can request a limited number of additional matches against higher elo opponents each day.
Can I prepare my bot before knowing the variant? +
Yes, you can do whatever you want before the competition, but you will likely have to make significant changes to work with the custom variant.
What resources are provided? +
We provide a complete development framework, documentation, example bots, and office hours with the organizing team. You'll also have access to our Discord community for support.
Leadership
Developers
Founders
Alumni
Contact Us
Address
- Carnegie Mellon University
- 5000 Forbes Avenue
- Pittsburgh, PA 15213
© 2025 CMU Data Science Club. All rights reserved.