TAI: You Learning Agent

TAI: You Learning Agent

TAI: You Learning Agent

03
TAI
UX Design
UX Design
AI Product
AI Product
{ SCROLL }
Intro

I worked with the Teaching Assistant Intelligence (TAI) team, a cross-functional research and product group at UC Berkeley’s Vive Center, focused on creating AI-powered tools to transform how students learn in large STEM classes. On this project, I collaborated with two other UI/UX designers, two product managers, and a team of five to ten engineers. I was responsible for the redefining the new user flow for the up-coming version, refine key features such as note, knowledge base, and file/video chat functions.

Trailer Video

MVP

Personalized, course-specific teaching assistant that supports students throughout their learning process

Personalized, course-specific teaching assistant that supports students throughout their learning process

Personalized, course-specific teaching assistant that supports students throughout their learning process

This is what longboard dancing looks like. Dancers usually post-edit music or listen to music while they dance.

Here’s what Jambo sounds like. The rider’s dancing style is captured by the board and transformed into sound.

The Problem

Lecture recordings are automatically broken down into meaningful segments, each paired with its transcript. Students can skim through the chaptered timeline or click on transcript sections to jump to specific moments, making long videos searchable and digestible.

Lecture recordings are automatically broken down into meaningful segments, each paired with its transcript. Students can skim through the chaptered timeline or click on transcript sections to jump to specific moments, making long videos searchable and digestible.

Lecture recordings are automatically broken down into meaningful segments, each paired with its transcript. Students can skim through the chaptered timeline or click on transcript sections to jump to specific moments, making long videos searchable and digestible.

Sample sound

Pitch bend

tempo

Sample sound

Pitch bend

tempo

When a course PDF is uploaded, the assistant automatically parses its structure—detecting sections, headings, and key concepts. These are presented as a sidebar table of contents and overview bullets inside the chat, giving students multiple ways to navigate. Students can jump directly to relevant sections or ask file-based questions, and the model responds with precise, section-grounded answers.

When a course PDF is uploaded, the assistant automatically parses its structure—detecting sections, headings, and key concepts. These are presented as a sidebar table of contents and overview bullets inside the chat, giving students multiple ways to navigate. Students can jump directly to relevant sections or ask file-based questions, and the model responds with precise, section-grounded answers.

When a course PDF is uploaded, the assistant automatically parses its structure—detecting sections, headings, and key concepts. These are presented as a sidebar table of contents and overview bullets inside the chat, giving students multiple ways to navigate. Students can jump directly to relevant sections or ask file-based questions, and the model responds with precise, section-grounded answers.

Force-sensitive resistor: sandwich the Velostat between two squares of copper tape.

Drum Pad: based on which part of the board is pressed, different samples will play.

Sample sound

Velostat

copper tape

copper tape

Force-sensitive resistor: sandwich the Velostat between two squares of copper tape.

Drum Pad: based on which part of the board is pressed, different samples will play.

Sample sound

Velostat

copper tape

copper tape

Force-sensitive resistor: sandwich the Velostat between two squares of copper tape.

Drum Pad: based on which part of the board is pressed, different samples will play.

Sample sound

Velostat

copper tape

copper tape

❌ Limited personalized attention: Office hours are overcrowded and intimidating.

🧩 Fragmented learning materials: Lecture slides, past exams, labs, and readings are scattered across different platforms, making it hard to study systematically.

🤖 Generic AI isn’t enough: Tools like ChatGPT can’t provide course-specific, reliable answers — leading to frustration or misinformation.

👉 This leads to students relying on guesswork, overloading TAs with repetitive questions, and missing deep conceptual understanding

❌ Limited personalized attention: Office hours are overcrowded and intimidating.

🧩 Fragmented learning materials: Lecture slides, past exams, labs, and readings are scattered across different platforms, making it hard to study systematically.

🤖 Generic AI isn’t enough: Tools like ChatGPT can’t provide course-specific, reliable answers — leading to frustration or misinformation.

👉 This leads to students relying on guesswork, overloading TAs with repetitive questions, and missing deep conceptual understanding

❌ Limited personalized attention: Office hours are overcrowded and intimidating.

🧩 Fragmented learning materials: Lecture slides, past exams, labs, and readings are scattered across different platforms, making it hard to study systematically.

🤖 Generic AI isn’t enough: Tools like ChatGPT can’t provide course-specific, reliable answers — leading to frustration or misinformation.

👉 This leads to students relying on guesswork, overloading TAs with repetitive questions, and missing deep conceptual understanding

User Background

User Background

Medium (Jambo)

Medium (Jambo)

Experience

Experience

Music Driven By

Emphasis on Dancing

Music Driven By

Emphasis on Dancing

Brings new perspective to

Music Production

Brings new perspective to

Music Production

Music Driven By

Emphasis on Musicianship

Music Driven By

Emphasis on Musicianship

Brings new perspective to

Longboard Dancing

Brings new perspective to

Longboard Dancing

Design Philosophy

From local "ChatGPT" to End-to-end Intelligent Learning Agent

Large STEM classes at Berkeley face three major pain points:

Additional micro adjustments & customization within Max/msp

potential for generative sound synthesis & effect

Gather data of dancer’s movements using Arduino

feet position, wheel speed, and board tilt

Map value to MIDI Control in Ableton Live

to sample sound, tempo, and pitch

layout

Jambo 1.0

Jambo 2.0

triggering multiple sensors all the time

clearance at the center avoids accidentally triggering sounds.

shape & size


the shape and size of the new sensor pad resemble a foot, which is more ergonomic for dancers.

Jambo 1.0

Jambo 2.0

when the dancer is simply pushing forward, the front foot stays near the center. This new layout makes music interaction more intentional.

fabrication

old sensors are unpredictable and too sensitive. we added more elastic material as a cushion in-between. New sensors are more reliable, stable, and durable.

Adhesive Foam Tape

FSR-touch sensor

silicone grid sheet

acrylic sheet

background beats


setting a rhythmic background

switch between multiple sets of samples

one-step

mix & switch

Try the MVP: tai.berkeley.edu

Try the MVP: tai.berkeley.edu

Try the MVP: tai.berkeley.edu

54 mm

32 mm