Case Study

Case Study

Case Study

Case Study

mia, a Machine Learning Platform

Overview

Overview

Overview

Overview

Overview

Data scientists specialize in creating machine learning models that can predict real-world data or make data-driven recommendations based on input data; but deploying these models require more than just science. Redapt reports that 90% of machine learning models never make it to production. The reason being is that majority of data scientists do not have the required skill set to deploy their models.

mia (Make Intelligent Applications) is an online platform that helps users to deploy their machine learning models into web applications simply and quickly via forms, eliminating the need for front-end development.

mia tasked our team to redesign their landing page and forms (uploading a model and app creation) so that mia can attract users and increase user engagement.

On top of their redesign goals, they wanted mia's brand to express fun and their app creation form to be reinvented into something more interactive (i.e. website builder).

Role

UX Designer

Applications

Focus Areas

User Research

Design Systems

Animation

Focus Areas

User Research

Design Systems

Animation

Focus Areas

User Research

Design Systems

Animation

Focus Areas

User Research

Design Systems

Animation

Focus Areas

User Research

Design Systems

Animation

Client Feedback

April 20, 2021

This has been an amazing experience. It’s amazing to see it all come to fruition. Congratulations you really did an amazing job with this.

Reply

Design System

Design System

Design System

Design System

Design System

Analyzing mia

We discovered that the user flow had at least 15 steps, which excludes the additional steps caused by potential errors and the number of inputs users need to add for their app.

Interviews

Interviews

Interviews

Interviews

Interviews

We carried out two rounds of interviews with 30 data scientists, the first one focusing on the usability of the website through contextual inquiry and the second one learning about mia’s users thoughts, frustrations, and motivations.

We created a persona, James the data scientist, and his journey with mia’s current website from our interviews’ key findings.

Key Opportunities

Key Opportunities

Key Opportunities

Key Opportunities

Key Opportunities

From our user's journey, these are our 4 key opportunities to solve our challenge:

Key Opportunity 1

Add more information to mia’s landing page

Our users wanted more information on how mia works from the landing page. The video and the short descriptions of its main three benefits were not enough to engage users.

We compared companies that offered similar services to mia and focused on how they presented their features.

From our competitive analysis, mia's landing page lacks detailed descriptions of their features, benefits to using their platform, testimonials from users or companies

  • Offers interactive visuals to show what the process would look like

  • Landing page has great visuals and exciting colors to entice users in learning more

  • Offers a systematic breakdown of what their product offers

  • Easy to follow headlines

  • Testimonials from customers working at big name companies

Efficiency Opportunity

Efficiency Opportunity

Efficiency Opportunity

Efficiency Opportunity

Efficiency Opportunity

Improve the efficiency of uploading your model

To make the form short and simple, we created a one-screen form where users could interact with the checklist of requirements, selecting their framework and version. The upload requirements would tailor to their selections. As a result, users would only read the information that pertains to their model.

Key Opportunity 3

Reinvent the app creation process

We moved away from mia’s traditional form design and redesigned the app creation process to be customizable “builder” experience.

Users can drag and drop inputs to the app form and edit the inputs on the side panel. Design offers more of a "Do-It-Yourself" feel.

We conducted A/B testing between this design and others, and as a result, 8 out of 8 users preferred the “builder” experience.

Before

After

Design System

Design System

Design System

Design System

Design System

Implement Design System

All interviews mentioned that mia’s platform lacked trustworthiness because of its UI flaws. I developed a design system to ensure consistency and professionalism throughout.

Color Palette

Color Palette

Color Palette

Color Palette

Color Palette

Typography

Typography

Typography

Typography

Typography

mia chose "Quicksand" as its primary font for its roundness. However, they dislike bolding the font for their headers due to its "comicy" look with its curved edges. As a result, users lack direction or CTA when viewing mia's content since all the headers and paragraphs and have the same weight.

I looked at other options and recommended to go with "Poppins" for the headers.

H1

H1 Heading

H1 Heading

H2

H2 Heading

H2 Heading

H3

H3 Heading

H3 Heading

H4

H4 Heading

H4 Heading

H5

H5 Heading

H5 Heading

H6

H6 Heading

H6 Heading

Grid System

Grid System

Grid System

Grid System

Grid System

To make sure components are properly aligned, I created a spacing guide.

Designs

Designs

Designs

Designs

Designs

** not all designs are shown

Upload a Model

Models

App Builder

You've reached the end!

You've reached the end!

You've reached the end!

You've reached the end!

Move on to the next case study.

You've reached the end!

Move on to the next case study.

© 2025 Lien Nguyen

© 2025 Lien Nguyen

© 2025 Lien Nguyen

© 2025 Lien Nguyen

© 2025 Lien Nguyen