Sage – UT Austin’s AI Teaching Assistant
Our team at UT Austin worked with Amazon Web Service on this on-going project.
We aimed to enhance student learning by developing a solution allowing faculty to build customized AI-powered chatbots tailored to their course’s content and pedagogical approach. The chatbots will leverage large language models to engage students in interactive and knowledge-building conversations based on materials provided by the faculties.
Team size

17 Team Members

 

Timeline

April 2024 – Now

 

 
 
Role

AI & UXUI Lead

 

Base

AUSTIN, TX, USA

 
Responsibilities

– Designed workflows for various user roles and created wireframes and style guides.

– Developed high-fidelity interactive interfaces with animations and conditional logic.

– Collaborated with engineers ,cloud architects and product managers for seamless integration with conversation design, backend LLM APIs, software development, and identity authentication.

Goal

Enhancing Personalized and Interactive Teaching

Data-Driven Insights for Improved Teaching Plans

Advancing Teaching Methodologies through AI

Reducing Repetitive Tasks for Educators

In collaboration with the AWS Generative AI Innovation Center (GenAIIC), UT Austin aims to develop a robust and secure POC for the faculty chatbot builder platform.

UT Austin plans to transition the solution to their internal team for moving to production, targeting a larger deployment in fall of 2024.

Target User

Challenge

Conversation and Content Design

How can Sage’s interactive capabilities effectively support course content? How can Socratic dialogue be employed to guide students in critical thinking and problem-solving?

 

Workflow and Interface Design 

How can we reduce the AI knowledge barrier and cognitive load, enabling instructors to train customized AI models? How can we simplify the process while ensuring comprehensive functionality tailored to course needs?

 

Privacy and Security

How can we ensure the necessary security for materials used across the university? How do we manage access permissions for different bots across various user roles?

Impact

Process

Demo

It’s an on-going project. And if you’d like to learn more or see the demo, feel free to reach out through email or LinkedIn. I am looking forward to sharing and connecting with you.

And SAGE is on its way : )