Milestone 3 : Midterm Presentation

These guidelines are meant to provide general direction for preparing your Milestone 3 presentation. Every project is unique, so if you believe your work doesn’t fully fit within these expectations, please discuss it with your TF. The goal is to ensure that your presentation aligns with the spirit of the assignment while still reflecting the specific goals and scope of your project.

Practical Information

  • Presentation Date: 10/28
  • Locations and Schdeule: Ed

Overview

In the past few weeks, we’ve explored the development of AI applications, focusing on:

  • Containerization using Docker.
  • Large Language Models (LLMs) — including instruction tuning, Retrieval-Augmented Generation (RAGs), Agents, and fine-tuning.
  • App Design, covering motivation, user flow, wireframing, solution and technical architecture, and code organization.
  • Serverless Deployment using Cloud Functions and Cloud Run.
  • ML Workflows with Vertex AI for scalable model training and automation.

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.

Presentation Requirements

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. Solution

    • Unique Value Proposition: Highlight why your solution solves the problem better than anyone else.

    • Demonstrate Value: Show clear, measurable benefits for your target users.

  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: Highlight how the solution can expand and adapt to new opportunities.

Be Prepared to Answer Questions On

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 and Competitive Advantage

  • What alternatives currently exist, and how does your solution outperform them? What evidence suggests real demand or potential for growth in this space?

Presentation Guidelines

  • Audience Consideration: Make your presentation accessible to both technical and non-technical stakeholders.
  • Engaging Narrative: Connect the technical details to real-world outcomes.
  • 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.

Deliverables

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.

Evaluation Criteria

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.
  • Engagement and Delivery (15%): Presentation style and ability to engage the audience, including during Q&A.
  • Visual Communication (10%): Use of visuals to enhance understanding.
  • Speaking Roles: Given the short time, not everyone needs to speak equally — but it’s a great opportunity for everyone to practice public speaking. Make sure each team member contributes meaningfully, even if briefly.
  • Follow Pavlos’ 10 Commandments for a Good Presentation — they capture what separates a good talk from a great one.

Template Repository: AC215 Milestone 3 Template Repository

Presentation Tips (Pavlos’ 10 Commandments for a Good Presentation)

  1. Keep slides concise — avoid overcrowding them with text.
  2. Rehearse your presentation multiple times to stay within the time limit.
  3. Anticipate questions and prepare thoughtful responses.
  4. Always include slide numbers for easy reference.
  5. Speak as a team — avoid phrasing like “I did this” or “I built that.”
  6. Avoid NPC (non-playable character) slides — every slide should have a purpose and move your story forward.
  7. If you look bored, your audience will be bored too — show energy and enthusiasm.
  8. Pace yourself — speak clearly, pause for emphasis, and let key points land. Avoid monotonous voice
  9. Think hard about “Future Work.” Show real understanding of the problem. Avoid vague lines like “get more data” or “make it scalable.” Be specific about next experiments, metrics, users, or failure modes you’ll tackle next.
  10. End strong — finish with a clear takeaway, not “That’s all.” Leave the audience knowing why your work matters.

Grading (25 points total)

  • 1 point - Peer and Self Evaluation
  • 4 points - Adjustment based on Peer, Self and TF Evaluation (contribution, meeting attendance)
  • 20 points - Final presentation, evaluated on:
    • Technical Depth
    • Business Acumen
    • Clarity and Structure
    • Visual Communication
    • Engagement and Delivery