Redesigning how issues and occurrences are escalated from autonomous vehicles to first responders for troubleshooting and resolution.
Learning
Interviewing & Shadowing AV operators and troubleforce.
I started with an extensive research process where I conducted user interviews and shadowed troubleforce (triaging team) during their work shift to meeting with different stakeholders to understand both the product and its problem space.
Through this, I was able to learn what their process is and how I can make their workflow better.
Research
Identifying and documenting painpoints/opportunities
The current flow from AVs (autonomous vehicles) to different areas for troubleshooting is complex. Each color represents a different team.
Not only do we have different users in the process, they also use many different tools and interfaces that could use some QOL updates. I synthesized the issues into a spreadsheet.
Final MVP
Allowing Troubleforce to be less dependent on other teams while triaging.
The original triaging and troubleshooting process required a lot of back and forth between different teams, which delayed the troubleshooting team. I proposed an improved flow through integrating different software and tools together to increase efficiency.
Occurrence Queue improvements
Troubleforce have to manually set the filters to view ongoing occurrences, and don’t use the “uncategorized” and “Work Orders” tabs.
There are 22 columns in the table to show all of the information about the occurrence but are not all used in this view.
Now, occurrences will be automatically sorted into “unresolved” and “all” occurrences. This will save time and make it very clear what occurrences are currently ongoing. Columns are cut, focusing on showing key information they need to know before clicking into the occurrence.
Current table
Proposed table
Information Architecture improvements on the detail page
When I was shadowing Troubleforce, I realized they wasted a lot of time scrolling up and down the page repeatedly to compare and stay updated to key information. To tackle this, we made a sticky section at top for info they need all the time, on top of cleaning up the UI to make it easier to scan.
Original details page
Updated details page
Slack integration UI improvements
Another thing I saw that could use small UI adjustments that could improve QOL for the team was to make the Slack Integration bot easier to read.
Original slack integration message UI
Proposed
Reflection
Interviewing and shadowing users’ process shows you all of the big and small opportunities that can make a difference.
When the project was first handed to me, it seemed straightforward – “Make the occurrence details page better.” But through actually interviewing and shadowing the different teams in an attempt to understand what it is that they do, was how I was able to figure out the different pain points and opportunities I could really help with.
Shoutout
Thank you to “SALSAS”
Without the most supportive, funny and kind teammates of the Fleet Design Team (SALSAS), my experience would’ve been extremely different.
Huge shoutout to my manager Stephanie who met with me every week to check on me, and helped me plan for the coming week. I definitely felt like any insecurity or questions I could always reach out for help!
Finally to my mentor Ashley who literally was there every step of the way, providing me with actionable feedback. While I knew I would get a mentor, I couldn’t have imagined how much I could learn from another person in such a short amount of time.
Designing Custom merch for the interns
WINTERNZ JACKET
As the only design intern, I felt I might not be able to connect with the other engineering interns. However, we were soon able to build relationships with each other, and even did a design project together to create our own unique swag!
What came after
After my successful internship, I went back to Cruise on a 9-month contract to join the ML/AI tooling team.
I worked on a machine learning/AI labeling tool to help offshore labelers more efficiently label driving environments to help our AV become smarter. Although that work is under NDA, I am happy to share more in an interview! Let’s chat :)
