layout | title | parent | nav_order |
---|---|---|---|
page |
Milestone 3 |
Projects |
3 |
In the past few weeks, we’ve explored the development of AI applications, focusing on:
- Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) techniques.
- Fine-tuning models for enhanced performance.
- Utilizing Google Cloud Platform (GCP) for cloud-based AI deployments.
- Managing data versioning with DVC and deploying applications using Docker containers.
- Employing Vertex AI for scalable model training and deployment.
For this milestone, you will present your AI application in the form of an investor pitch. Your objective is to communicate complex technical concepts clearly and persuasively while demonstrating the potential business impact of your AI solution. Your presentation should:
- Presentation Date: Mar 27
Details on the schedule will be posted on Ed.
Slides submission: Please submit slides by noon on 3/27 via Canvas.
Your presentation should be 5 minutes long, followed by 1 minute for Q&A. It should include:
1. Problem Statement and Target Audience
- Define the Problem: Clearly articulate the specific problem your AI application addresses.
- Identify the Target Audience: Describe who will benefit from your solution (e.g., specific industries, user groups).
2. Unique Value Proposition
- Highlight Uniqueness: Explain what sets your solution apart from existing alternatives.
- Demonstrate Value: Illustrate how your AI application provides unique benefits to the target audience.
3. Scalability and Efficiency
- Technical Scalability: Discuss how your application can scale to meet growing demands.
- Performance Optimization: Explain any optimizations implemented for efficiency (e.g., fine-tuning, infrastructure choices).
- Infrastructure Considerations: Briefly mention the technologies used (e.g., GCP, Docker, Vertex AI) and why they were chosen.
4. Future Development and Growth Potential
- Next Steps: Outline your plans for further development of the AI application.
- Market Growth: Discuss the potential for market expansion and how your solution can adapt.
- Technical Architecture and Infrastructure Choices Why did you choose specific tools and platforms? How does your architecture support scalability and reliability?
- Data Management and Security How do you handle data versioning and storage? What measures are in place to ensure data security and compliance (if applies)?
- Model Performance, Optimization, and Maintenance What are the performance metrics of your model? How will you maintain and update the model over time?
- Market Opportunity What is the size of the market you are targeting? How does your solution meet a demand or gap in the market?
- Audience Consideration: Tailor your presentation to be accessible to both technical and non-technical stakeholders.
- Engaging Narrative: Tell a compelling story that connects the technical aspects to real-world impact.
- Visual Aids: Use clear and effective visuals (e.g., diagrams, charts) to illustrate key points.
- Time Management: Practice to ensure you cover all points within the 5-minute timeframe.
Note: All deliverables must be submitted via GitHub (milestone3 branch), submit full commit hash on Canvas.
1. Presentation Slides
2. Code submission, similar to Milestone 2 Any additions/modifications must be highlighted in the README.md file.
Your presentation will be assessed based on:
- Technical Depth (30%): Clarity and accuracy in explaining your AI application’s technical aspects.
- Business Acumen (25%): Effectiveness in conveying the market potential and value proposition.
- Clarity and Structure (20%): Organization of content and logical flow of the presentation.
- Visual Communication (15%): Use of visuals to enhance understanding.
- Engagement and Delivery (10%): Presentation style and ability to engage the audience, including during Q&A.
-
Template Repository: E115 Milestone 3 Template Repository
-
Presentation Tips:
- Keep slides concise; avoid overcrowding with text.
- Rehearse your presentation multiple times.
- Anticipate questions and prepare responses.