Methodology & Results

We will deliver two websites: the students and teacher platform. Our methodology covers the technical stack used, development phases, and testing procedures. Results will showcase both platforms and feedback from initial user trials of both websites.

Methodology

This methodology allows for flexibility and continuous improvement based on internal and external feedback from various technical stakeholders.

Requirements Gathering and Project Initiation

We will have an initial meeting with the project’s stakeholder to confirm detailed requirements and specifications based on the needs of students, teachers, and the school. Furthermore, we will also lay down the comprehensive project plan with clear milestones and intended deliverables.

Research

We will conduct thorough research on existing RAG (Retrieval-Augmented Generation) solutions and best practices within the educational technology sector.

1. Literature Review

  • Analyzing academic papers on Retrieval-Augmented Generation (RAG) and large language models
  • Studying text chunking methods and prompt engineering techniques

2. Technology Stack Assessment

  • Comparing embedding models for knowledge representation
  • Assessing vector databases for efficient information retrieval
  • Exploring Local LLM options suitable for our resource constraints

Development

Building up on top of the extensive research conducted, we will then focus on development of our project.

1. Version Control and Collaboration

  • Using Git and GitHub for version control and project management
  • Implementing pull requests and code reviews for quality assurance

2. Tools and Technologies

  • Frontend: Next.js with TailwindCSS
  • Backend: FastAPI
  • Database: PostgreSQL
  • Cloud: AWS (EC2, S3, ECS)
  • AI Model Deployment: Lightning AI

3. Design and Implementation

  • Preprocessing data sources for accurate retrieval
  • Developing a modular AI tutor system
  • Creating core LLM-RAG chat system
  • Building exercise and report generation features

4. Testing and Quality Assurance

  • Conducting alpha and beta tests with students and teachers
  • Gathering feedback for system improvement

5. Security and Data Protection

  • Implementing data minimization and encryption
  • Using AWS security services (KMS, CloudTrail, AWS Config) for data protection

Assessment

Once the product has been successfully developed and passed all the tests outlined in the above, we will focus on assessing the impact as well as various documentation to help support teachers and students navigate through the application.

1. Documentation and Training

  • Creating user manuals and technical guides
  • Developing training materials for teachers and students

2. Evaluation and Reporting

  • Establishing key performance indicators (KPIs)
  • Regularly assessing system performance
  • Conducting a final project evaluation

ℹ️

To view a more comprehensive description of our methodology, please take a look at our project plan.

Results

We will showcase the two websites: the students and teacher platform, alongside the feedback obtained from initial users of the two platforms.

Coming soon…