-
Notifications
You must be signed in to change notification settings - Fork 1
Project Milestones
By the end of Term 1, the team will establish the core infrastructure required for AI development. This includes setting up Gradio, Chroma, Ollama, and LangChain, ensuring a seamless foundation for building AI chatbots and AI-driven systems. The goal is to have a fully functional development environment that allows for rapid prototyping and deployment of AI solutions starting from Term 2.
Current State: Done
Purpose: Deliver a functional UI that lets users interact with the AI and reflects stakeholder expectations.
- Use Gradio for immediate UI development; Open WebUI for future transition.
- Brand the UI ("Model Works"), accept user prompts, and display AI responses.
- Gather feedback and refine the UI iteratively.
- Success: Document via screenshots; achieve stakeholder satisfaction.
Current State: Done
Purpose: Develop scripts and orchestration tools to enable modular deployment of AI models and UI layouts. This is very important because it is a core functionality to allow the JoeyLLM to connect to ModelWorks infrastructure.
- Script to toggle UI layouts via file paths.
- Set up Docker orchestration using Docker Compose to manage services easily.
- Success: Proven ability to interchange components; tested for flexibility and smooth container orchestration.
Current State: Done
Purpose: Make the AI chatbot accessible via the public-facing Model Works website.
- Integrate the developed UI with the Model Works web frontend.
- Ensure the chatbot is live and usable through the website's main domain or subpath.
- Set up secure access and verify functionality in production.
- Success: Users can interact with the chatbot directly from the website.
Current State: Done
Purpose: This model switching feature is directly required for JoeyLLM (when JoeyLLM has an API of some sort to connect).
- AI container script for simple model switching (e.g., Ollama models).
- Allows us to switch models based on confidence score.
Current State: To do
Purpose: Enable the application to retrieve information from a provided link in real time, delivering relevant data to users.
- Integrate the ability to pull information from a provided link.
- Ensure real-time retrieval and presentation of relevant data.
-
Success: Demonstrate the system fetching data accurately from the link.
Current Status: Done
Purpose: Enable the model to remember and store both AI and user conversations for future reference.
- Store conversation history as a list.
- Ensure the model recalls both AI and user interactions.
-
Success: Verify the model can reference previous conversations accurately.
Current Status: To do
Purpose: Enable agents to interpret user input, execute predefined functions, and integrate with external services.
- Agents interpret user input into designed/predefined function calls.
- Explore integration with other software/services (e.g., Google Calendar, Drive).
- Develop chatbots and agents for enhanced workflows.
-
Success: Demonstrate agents performing tasks and integrating with external services.
Current Status: Done
Purpose: Create reusable AI pipeline components and templates to streamline development across multiple projects.
- Write classes implementing AI pipelines reusable in multiple settings.
- Develop common components to be frequently used across projects.
-
Success: Demonstrate reusability of components in different AI workflows.
Current Status: Done
Purpose:
Establish a stable connection between the Model Works middleware and the Joey LLM API to enable end-to-end demo workflows.
- Finalize and adapt middleware request/response handling to match Joey LLM API format.
- Implement integration layer to connect front-end prompts through middleware to the Joey LLM.
- Prepare and run integration test cases to validate prompt–response flow.
- Document the integration process for future reference.
Success:
Front-end demos run seamlessly through the Joey LLM API without relying on internal AI application.
Current Status:
In Progress
Purpose:
Deliver final showcase materials — a poster, script, and video — to present the project’s progress and outcomes over the year.
- Design and finalize the project poster (due October 10, 5:00pm) using tools such as Canva or Miro.
- Draft and refine a script to clearly communicate the 18 weeks of project progress.
- Record and edit a video presentation based on the script.
- Ensure all materials align with showcase requirements and are easy for observers to follow.
Success:
Poster submitted by the deadline, and both the script and video are completed and ready to present at the showcase.
Current Status:
In progress
Welcome to the ModelWorks Wiki! Use the links below to navigate our resources quickly.
Product Information
- Procedure to Resolve Conflicts
- Project Milestones
- Decision Making Protocol
- Story Point Estimation Algorithm
- Gradio Notes
- Javascript Notes (split into two)
- Ollama Notes
- Langchain Notes
- Chroma Notes
- GPU Monitor Notes
- Open WebUI Notes
- Web Search Notes
- Chat History Notes
- Meeting Minute 1
- Meeting Minute 2
- Meeting Minute 3
- Meeting Minute 4
- Meeting Minute 5
- Meeting Minute 6
- Meeting Minute 7
- Meeting Minute 8
- Meeting Minute 9
- Meeting Minute 10
- Meeting Minute 11