Project
Guidelines for Final Project Presentation
You will work as a team to design, develop, and deploy an AI-powered business solution that leverages large language models (LLMs) and generative AI technologies. This project is a collaborative effort, where each team will consist of up to six members. The team will work together to identify a real-world business problem, propose an innovative solution, and implement it using the tools and concepts covered in this course.
At the outset of the project, the team will submit a Group Charter that outlines the roles and responsibilities of each member. This document should clearly define individual contributions, such as project management, model development, API integration, interface design, testing, and deployment. The Group Charter ensures accountability and fosters effective collaboration throughout the project.
The team will use GitHub or the internal GitLab repository to manage their project codebase. This includes version control, collaboration on code development, and documentation. Instructors and TAs must be granted access to the repository from the beginning to provide guidance and support. While teams may choose to keep their repositories private during development, this project serves as an excellent opportunity to create a public portfolio piece for showcasing your skills in job interviews or professional settings.
The AI application developed by the team must include the following components:
- A Web Interface: The application should have a user-friendly web interface that allows users to interact with the AI system. This could be implemented using frameworks like Flask, FastAPI, or Streamlit. Alternatively, no-code platforms such as Dify, n8n, or Amazon PartyRock may also be used to build the interface and manage workflows.
- AI Functionality: The solution must incorporate LLM capabilities using OpenAI or Ollama APIs. Teams are encouraged to explore advanced features such as fine-tuning models, chaining prompts, integrating external data sources, or utilizing no-code tools for model orchestration.
- Deployment: The application must be hosted internally (e.g., on a server or cloud platform) or publicly to ensure accessibility for testing and review. The deployed version should allow classmates to interact with the demo for evaluation purposes.
The team will also document their development process thoroughly by maintaining a README file in their repository. This documentation should include:
- A description of the business problem being addressed.
- An overview of the solution architecture.
- Instructions for setting up and running the application.
- Details on how LLMs were used and any challenges encountered during development.
- If no-code tools were used, an explanation of how they were integrated into the solution.
Finally, selected team will present their project during a demo session at the end of the course. The presentation should showcase the functionality of the application, highlight key technical achievements (including creative uses of no-code platforms if applicable), and reflect on lessons learned during development. This assignment not only strengthens teamwork but also equips you with practical experience in building AI solutions and creates a tangible artifact for future professional use.
Evaluation
| Criteria | Exemplary (4) | Proficient (3) | Developing (2) | Beginning (1) |
|---|---|---|---|---|
| Problem Definition | Clearly identifies a real-world business problem and provides an innovative, well-researched solution. | Identifies a relevant business problem and proposes a feasible solution with some innovation. | Defines a general problem but lacks clarity or depth in the proposed solution. | Problem definition is vague or disconnected from a practical business context. |
| Technical Implementation | Demonstrates advanced use of LLMs, APIs, and/or no-code tools; solution is robust and well-optimized. | Effectively uses LLMs, APIs, and/or no-code tools; solution is functional with minor optimization issues. | Basic use of LLMs, APIs, or no-code tools; solution has functional gaps or inefficiencies. | Minimal or incorrect use of LLMs, APIs, or no-code tools; solution is incomplete or non-functional. |
| Web Interface | User-friendly, visually appealing interface with seamless functionality and responsiveness. | Functional interface with good usability and design but minor inconsistencies. | Interface is functional but lacks usability or visual appeal; may have significant flaws. | Interface is poorly designed, difficult to use, or non-functional. |
| Team Collaboration | Clear roles and responsibilities; excellent teamwork and communication throughout the project. | Defined roles and good teamwork; occasional communication issues but resolved effectively. | Roles are somewhat unclear; teamwork was inconsistent with some communication challenges. | Roles are undefined; poor teamwork and lack of communication hindered project progress. |
| Deployment | Fully deployed application accessible internally/publicly; performs reliably under testing. | Deployed application accessible internally/publicly; minor performance issues during testing. | Partially deployed application with limited accessibility or significant performance issues. | Application not deployed or inaccessible for testing/review purposes. |
| Documentation | Comprehensive documentation covering all aspects of the project, including setup instructions and challenges faced. | Sufficient documentation covering most aspects of the project with minor gaps in detail or clarity. | Basic documentation provided but lacks depth or clarity in explaining key aspects of the project. | Minimal or missing documentation; difficult to understand or replicate the project setup. |
| Innovation & Creativity | Highly creative and innovative solution that demonstrates original thinking and problem-solving. | Creative solution with some original ideas that address the business problem effectively. | Solution shows limited creativity and relies on standard approaches without much innovation. | Solution lacks creativity and originality; does not effectively address the identified problem. |
Supporting Documents
| Title | Description |
|---|---|
| Final Project Presentation | Follow these guidelines for your final project presentation |
| Project Ideas | A variety of innovative AI business solutions leveraging large language models |
| AI for Data Analytics | AI tools and techniques can enhance every stage of the data analytics pipeline, from data collection and cleaning to advanced analysis and visualization |
| Integrate OpenAI and Ollama API | How to use API calls to OpenAI, Ollama, and other services in your project |
| No-code Tools | No-code environments like Amazon PartyRock, Dify, and n8n to build AI-powered solutions without requiring extensive programming knowledge |