#Local AI
#Dev Tools
#CLI
#B2B
#Internship
Azure Foundry Local is an on-device AI inference solution offering performance, privacy, customization, and cost advantages for AI model deployment. It is scalable across 1.4 billion global users.
As the only UX Design & Research intern, I led a user landscape & experience research program to validate the user needs and the current roadmap. In addition I proposed 30+ product recommendations & design mockups to the executive VP team.
I also went beyond my expected tasks and took on extra design responsibilities for an internal AI Agent Building Platform including providing design system accessibility guidelines for multiple components, and providing interface design proposals.
Date
May 2025 - August 2025
Tools
Figma, Green Room, Microsoft Co-Pilot
Team
x1 VP
x5 Program Managers
x10 Software Engineers
x1 Technical Program Manager
x1 Product Marketing Manager
x1 UX Design & Research Intern (Me)
Context
What is Foundry Local?
Foundry Local is an end-to-end local AI solution for shipping applications that run entirely on the user's device. It provides a lightweight package including an easy-to-use SDK, automatic hardware acceleration, and optimized models. Users can deploy it on their local CLI.
No Latency
AI runs in real time without cloud latency
Save Cost
No expensive GPU or Cloud token costs
Full Privacy
All data stays on your personal device
Offline Access
Access anytime anywhere without internet connection
Process
Research Program Planning
Objective
Understand the user needs, pain points, and overall experience with developing & deploying AI models locally
Obtain feedback on product direction of Azure Foundry Local
Competitive analysis across competitor products
Method
1:1 Interviews : 12 Microsoft Azure customers & developers
Usability Tests & Cognitive Walkthrough : [Foundry Local Build 0.4.91 Release]
Main Research Areas
Interviews & Usability Tests
Interviews & Usability Tests with Microsoft customers
Interviewees include AI developers, CEOs, Mobile AI app developers and PhD researchers.
Each 1hour interview includes a user cognitive walkthrough of the most recent public build of Foundry Local, to observe potential obstacles throughout the experience.
Insights
Where we were aligned with the users
Developers appreciate the Microsoft Ecosystem Integration
Seamless flow with VS Code AI Toolkit, AI Dev Gallery, and other Microsoft tools
Simplifies managing models and experimentation
Can replace Ollama leveraging Microsoft’s support and integration
For flexibility, compatibility, and fits team internal skillsets / client device ecosystem
iOS has platform restrictions and loading issues
Many developers use open-source Hugging Face models for experimentation & development
Need an efficient way to manage fine-tuned models between the local environment and Hugging Face.
No other platform supports Agents
Solutions are needed for significant amount of Agent use cases in local environments
Healthcare, Law enforcement, Defense, Finance industries, Labs
Where we were losing users
The obstacles that users face while using Foundry Local could be broken down into 3 levels.
Entry Level : the phase before users actively use Foundry Local, as they set up their environment and learn about the pre-requisites
Development Level : the phase where users actively use Foundry Local
Production Level : the phase where users finish development and prepare for product shipping
Product Recommendations
Based on the insights gathered, I presented a slide deck of 30+ product recommendations & 5+ CLI Design Mockups.
Feedback & Outcome
Feedback
After presenting my research findings, recommendations, and design proposals to 2 VP executives & teams,
I gained positive reviews and feedback from my manager and stakeholders whom I have worked with over the internship.
Outcome
The Computational Design Team of AI Platforms lead by John Maeda was the best team I have ever had the honor of working with. They were incredibly welcoming, humorous, and supportive throughout the internship allowing me to learn not just professional skills as a researcher & designer, but to be a team player in a cross-functional setting.
Copyright©2026. All right reserved Sun-Q Kim