Note: This blog post is for testing and context-setting. It is deliberately brief and should not be considered the official first post for grading purposes.
Hello, world! I'm so excited to get started with this project, and this key step is getting the blog up and running.
My capstone project is designing and implementing an agentic team engineering system. This system will let users design teams of existing or custom AI agents to accomplish tasks like creating software, running and debugging code, and even more basic tasks like helping manage your inbox. In a nutshell, I'm hoping to implement a software that runs locally (as in, is based locally but works with cloud models) to accomplish tasks that one would use Claude Code or Codex for today.
Looking ahead, the deliverable should be
- A well-packaged, locally-hosted system that can be easily downloaded and run by users
- An environment that facilitates the building, saving, and running of teams and/or workflows of agents
- Highly customizable, i.e. the user should be able to use cloud or local models, create or import/add agents, and scale the software to his or her use case and hardware
I hope to build out the software in time to experiment with different agentic teams and see how they compete given the same specifications. Additionally, I'd like to design a GUI. I'll do my best to implement it as time allows, with the understanding that this is just a one-semester project.