PopK - The Future of Esports Ranking Inspiration We saw the need for a more transparent, objective and data-driven approach to ranking esports teams. Current systems are often subjective or don't reflect the full context of team performance.
What it does PopK introduces an innovative new ranking system for League of Legends using advanced machine learning and metrics. It trains AI bots to generate huge datasets, then analyzes multiple factors like win rate, KDA, objective participation to score teams.
How we built it We developed a scalable microservices architecture powered by Docker, Kubernetes, AWS and OpenAI. Bots are trained via reinforcement learning to automatically play matches. Frontend displays rankings and deep stats visualized with React.
Challenges we ran into Integrating with the LoL API, modeling complex player/team interactions, and deploying at global scale on AWS services. Ensuring bots exhibited human-like gameplay was also difficult.
Accomplishments that we're proud of Training over 150 bots to generate petabytes of competitive data. Developing 8 unique ranking metrics to analyze multiple facets of team performance. Building a fully functional MVP in under 6 months.
What we learned How to build large distributed systems, leverage AWS, and apply ML/AI techniques to esports. The value of an iterative development approach to complex problems.
What's next for PopK Expand bot training with newer games. Monetize through features like fantasy leagues, betting, and premium analytics. Ultimately, develop the most objective global esports ranking system.
Built With
- amazon-web-services
- aws-services-for-scalability/hosting
- bots
- database
- developed
- docker
- docker/kubernetes-for-deployment
- dynamodb
- dynamodb-for-database
- ec2
- elasticsearch
- elasticsearch-databases:-dynamodb-apis:-riot-games-api-(for-league-of-legends-data)
- frontend
- functions
- games
- here-are-the-main-technologies-used-by-popk:-languages:-javascript
- hosting
- indexing/search
- integration
- kubernetes
- kubernetes-cloud-services:-aws-ec2
- lambda
- microservices
- node.js
- nosql
- openai
- openai-api-(for-bot-training)-technologies:-microservices-architecture
- openai-for-ai/bot-training
- orchestration
- python
- python-(for-bot-training)-frameworks:-node.js
- react
- react-platforms:-docker
- riot
- scalable
- serverless
Log in or sign up for Devpost to join the conversation.