AI Gamer Project - Part 1 - Introduction¶
So I decided that I'm going to take a dive into AI...
I dabbled in it with the audio generation project for these blogs but I feel like that was a minimal use. So I'm using AI to help me learn AI. I'll be tackling this project in phases and will be documenting live!
Phase 1: Infrastructure Setup¶
Kubernetes on WSL – Set up a local Kubernetes cluster on your Windows machine using WSL. Minikube or K3s would be good lightweight options.
Deploy Grafana & Prometheus – Set up a monitoring stack within the cluster to collect and visualize metrics.
Phase 2: Game & AI Integration¶
Choose a Game – Select a game with a suitable interface for AI control. Something with a simple API or OpenAI Gym support would be ideal.
Agentic AI Model – Start with a reinforcement learning (RL) model (e.g., PPO or DQN) using frameworks like Stable-Baselines3 or RLlib.
Capture Gameplay & Model Actions – Use OBS Studio (or a custom GStreamer setup) to stream the game and AI model decisions.
Phase 3: Website & Dashboard¶
Live Streaming – Set up a video feed on your website (e.g., via RTMP to a web player).
Metrics Dashboard – Expose model training insights and actions on Grafana, embedded below the stream.
Would you like to start with setting up the Kubernetes cluster or the AI model?
Conclusion¶
Wish me luck!!! This is going to be one hell of an adventure!