Digital Learning Strategy AI Assistant Prototype to Support Educators
Project phases
Phase 1: Enhancing Digital Learning & Engagement with Generative AI
The Ministry of Post-Secondary Education and Future Skills (PSFS), in collaboration with the UBC Cloud Innovation Centre (CIC), has enhanced the AI-powered Digital Learning Strategy (DLS) AI Assistant Prototype to better support educators in post-secondary education.
The project took shape in two phases. Phase 1 resulted in an AI-prototype that would allow the post-secondary community to retrieve information related to the B.C. Post-Secondary Digital Learning Strategy (DLS) based on conversational queries. Phase 2 expands the prototype with added functionality for educators to upload their course documents and receive recommendations on how to align their course materials with the DLS.
Phase 1 helps students, the public, and the post-secondary community and Phase 2 builds on this foundation by introducing new functionality designed to support post-secondary educators.
Approach
In Phase 2, the UBC CIC leveraged Generative AI (GenAI) and Retrieval-Augmented Generation (RAG), to add functionality that allows for more personalized recommendations, deeper engagement, and more effective support for educators integrating the DLS. Educators can upload course materials and content, ask specific questions, and receive AI-generated guidance aligned with the DLS and its key supporting documents. RAG is used to retrieve the most relevant information from a pre-indexed vector database before generating responses, ensuring that answers are context-aware and aligned with DLS recommendations and supporting documents. The assistant prototype can process more complex queries and provide structured responses, improving its ability to support educators.
Supporting Artifacts
Educator View
Administrator View
Architecture Diagram
Technical Details
The prototype is fully deployable as infrastructure as code. You can deploy the solution by following the Deployment Guide available on the UBC CIC GitHub Repo.
Uploading and Managing Course Materials
Admins upload course materials to the application using a pre-signed upload URL, which are then stored in an S3 bucket. When a new Digital Strategy Assistant file is added to the bucket, it triggers a data ingestion workflow powered by AWS Lambda. This Lambda function processes the file, embedding the text into vectors using the Amazon Titan Text Embeddings V2 model via Amazon Bedrock.
Data Processing and Embedding
The generated vectors are stored in a PostgreSQL database for easy retrieval. Admins can manage Digital Strategy Assistant content by sending API requests that trigger Lambda functions, which interact with Amazon RDS to update and retrieve data.
AI Conversations with the LLM
Users can interact with a Large Language Model (LLM) Llama 3 by sending API requests that trigger another Lambda function to generate responses. The function runs a Docker container on Amazon ECR, storing conversation history in DynamoDB. Using Retrieval-Augmented Generation (RAG) architecture, the LLM pulls relevant course data from Amazon RDS to provide contextual answers.
Document Evaluation
Users can evaluate their documents against the DLS areas of discovery by uploading their documents which triggers a Lambda function to generate the embeddings for the document. The embeddings are then used by another Lambda which uses Retrieval-Augmented Generation (RAG) architecture, to pull relevant course data from Amazon RDS to the user with a detailed response about how their documents align with the areas of discovery.
Frontend and Security
By integrating AWS Amplify, API Gateway, AWS Lambda, and Amazon Bedrock, the system empowers administrators to efficiently manage course materials while enabling users to engage in intelligent, context-aware conversations powered by a large language model (LLM). The frontend, hosted on AWS Amplify, routes all user requests through API Gateway, which securely triggers the appropriate Lambda endpoints. All API endpoints are encrypted to ensure secure data transmission, and a built-in rate limiter protects the system by blocking excessive requests from individual IP addresses.
Link to solution on GitHub: https://github.com/UBC-CIC/Digital-Strategy-Assistant
Acknowledgements
Project Sponsors: Ministry of Post-Secondary Education and Future Skills
Student team: Harshinee Sriram, Kanish Khanna, Amy Cao
Materials from the B.C. Open Collection with a CC BY 4.0 license were used, with minor modifications to document descriptions for the purpose of testing the prototype: Environmental Geology Course Package by Steven Earle (Thompson Rivers University), © 2022 Steven Earle; Business Communications Course Package by Arley Cruthers, John Grant, Melissa Ashman, Petti Fong and Seanna Takacs; and Introduction to Management Course Package by Scott van Dyk (Coast Mountain College).
Photo by: Fauxels
About the University of British Columbia Cloud Innovation Centre (UBC CIC)
The UBC CIC is a public-private collaboration between UBC and Amazon Web Services (AWS). A CIC identifies digital transformation challenges, the problems or opportunities that matter to the community, and provides subject matter expertise and CIC leadership.
Using Amazon’s innovation methodology, dedicated UBC and AWS CIC staff work with students, staff and faculty, as well as community, government or not-for-profit organizations to define challenges, to engage with subject matter experts, to identify a solution, and to build a Proof of Concept (PoC). Through co-op and work-integrated learning, students also have an opportunity to learn new skills which they will later be able to apply in the workforce.
