I spent the summer of 2022 as a UX Design Intern at NVIDIA, with the NVIDIA GPU Cloud (NGC) team! I had the amazing opportunity to work cross-functionally with other UX Designers, UX Researchers, Engineers, and Product Managers to design new features for Fleet Command, NVIDIA's cloud-based platform for deploying and managing AI applications on edge devices. Although this project is under NDA, I would love to share my process and learnings from my 12 weeks at NVIDIA.
Please email me for more details about the work I did at NVIDIA!
(June - September 2022)
NGC UX Team
Laura Hanssen (Manager)
Krissel Norris (Mentor)
Hsiao-Han Tseng (UX Designer)
Adam Bronsther (UX Researcher)
My team, underwater! :)
Owned the end-to-end designs of a new device adding/pairing flow for Fleet Command, from concept to implementation
Designed an information architecture concept to integrate new design work with existing features on Fleet Command, and improve overall ease-of-use
Created and prototyped 10+ dashboard and landing page concepts for future production
Rapidly iterated wireframes based on feedback from design team, engineers, and product managers
Analyzed insights from usability testing to identify opportunities for UX improvements
Read user documentation and PRDs (product requirements documents) to understand business goals for Fleet Command
Presented designs weekly to stakeholders to validate features and provide rationale for design decisions
Created design specs to document edge cases for development
Validated designs and handed off final prototype to be implemented
Utilized Kaizen UI design system to maintain consistency and align designs with existing patterns
What is Fleet Command?
Fleet Command is a cloud platform that allows enterprise customers to manage, deploy, and scale AI applications on the devices that power their businesses. Fleet Command operates at the edge, meaning that the applications and the devices they are deployed to involve data processing, and this processing happens as close to the original data source as possible. At NVIDIA, where edge computing is combined with machine learning and AI software, this technology is called edge AI. Fleet Command and edge AI have many critical use cases, including managing traffic cameras, security surveillance systems, product inspection, quality control, hospital patient monitoring, and so much more.
Main Project Summary
The process of adding and pairing devices on Fleet Command is complicated and unclear. The three key steps: add a device, flash the system installer, and enter an activation code, are overly complicated. Managing devices is a key use case of Fleet Command, and users are frustrated by the amount of work required for a seemingly simple step.
I designed an end-to-end user flow with a simplified process for adding and pairing devices. I created over 20 iterations and presented them weekly to a team of stakeholders. I also collaborated closely with engineers and product managers to validate and defend my design decisions, understand technical constraints, and consolidate my work into a high-fidelity prototype for handoff.
My designs are currently being implemented by the NGC engineering team!
By streamlining the add/pair device flow, Fleet Command's users will have a clear, three-step process to easily set up their edge sites, and the flexibility to complete these steps when they want to. With this simplified setup, users will be able to start deploying new AI software more quickly to their edge devices.
A sneak peek of my process
Completing onboarding tasks
Meeting my team
Reading product documentation
Understanding the Problem
Exploring alternate solutions
Designing new features
Learning about Fleet Command
Analyzing usability findings
Conducting competitor analysis
Research and Ideation
Creating low-fidelity designs
Gathering early feedback
Narrowing scope and solution
Utilizing design system
Aligning designs with constraints
Implementing stakeholder feedback
Validation and Feedback
Creating design specs
Writing design documentation
Designing for edge cases
Final review of designs
Handoff meeting with designers, engineers, and product managers
1. Think Blue Sky
Before beginning the design process, I initially thought that there was only one possible solution for the problem. This caused me to become attached to certain ideas early in the design process and narrow my scope before exploring all possibilities. However, my mentor reminded me that, because I was driving this project, I had an opportunity to dive deeper and embrace a blue sky mentality. This allowed me to design concepts that were outside the box and rule out lots of different ideas, resulting in the best possible solution!
2. Be Bold!
As an intern on a team of very experienced designers and engineers, working on a highly technical product, there were times when I felt like I was in over my head. Pushing myself out of my comfort zone was challenging, but it gave me the courage to involve my teammates in every step of the process. I made the most of my coworkers’ experience by regularly seeking out and implementing their feedback, leading sync meetings to showcase and defend my work, and asking questions to learn more about Fleet Command.
3. Measuring Success
If I had more time, I would have liked to conduct usability testing to measure the efficacy of my designs and inform future improvements. After implementation, I would have measured the task success rate and task completion time for adding and pairing devices. I would also measure overall customer satisfaction and compare the average number of devices added per user before and after my solution was implemented.