WildCity: An AV’s Digital Proving Ground
September 23, 2026
May Mobility's Senior Director of Autonomy Data, Jonathan Voigt, shares how he uses data to solve problems and improve autonomous driving.

As a Senior Director of Autonomy Data, my teams ensure that the data generated by May vehicles gets to the systems and people that need it as quickly as possible. This delivery can be in real-time to power fleet operations and remote assistance applications, or in bulk after vehicles return for advanced analytics, triage, and debugging efforts by engineering. We also feed this data into the training and testing systems that improve the models driving the car, continuously improving driving behavior. We are also responsible for running the sites and datacenters that power our operations.
AVs collect an enormous amount of data and then use it to improve themselves. The challenge is collecting as much of that data as possible, and then transforming and organizing it - all while keeping costs down. We don’t have a blank check to spend here, unlike some of our competitors, so May prioritizes efficiency and sustainability. It is a hard, interesting problem that demands creative solutions every day.
A challenge I’ve faced at every startup, including this one, is getting more done with a small team. There is always an infinite amount of work to be done. The tools for AI-assisted engineering we use here have enabled me to personally contribute in a much more meaningful way - especially with small, important but potentially distracting tasks - and keep my teams focused on our most pressing issues.
I am proud of the team that I have built - the people I have hired and retained. Even as AI enables new ways to work, the people on the team are what matter most and are critical to how we build a great product.
Our existing systems were built by teams working under different constraints and on a smaller scale than we have today (and without the technologies available today). We are modernizing the stack with the benefit of a working example to compare to. And we are executing faster than ever. The wins are already rolling in.
Drive with purpose. I enable my teams and trust them to perform. We have an opportunity in front of us; it’s up to us to execute and show the world what we’re made of.
I’d love to see intelligence in the swarm, vehicles communicating with each other, coordinating to make traffic flow like water. Unfortunately, the coordination required across manufacturers and municipalities - not to mention the reality of trust - means this will likely remain science fiction for a long time.
I want to see everyone get the time back that we spend moving ourselves around. To use that time in a way that is best for us, maybe just to slow down and appreciate the world as it rolls by.
If you’ve never deployed to hardware before, it is both a wonderful and frustrating experience. We name our vehicles because they reveal personalities to us - something cloud software doesn’t do in the same way. Unlike software in the cloud, you can't easily tear down or replicate a vehicle. Watching software you've deployed to a vehicle drive away from you is something special.
Take a deeper dive into data with Jonathan Voigt in this episode of Autonomy Talks to learn what it takes to build a strong data foundation, why creative and cost-conscious problem solving matters, and how blending different approaches helps our vehicles handle the toughest moments on the road.


Our vehicles drive themselves, but it’s people like you who drive May Mobility. Join ranks with our diverse group of curious, passionate people and help us solve a spectrum of autonomous mobility challenges.
Our vehicles drive themselves, but it’s people like you who drive May Mobility. Join ranks with our diverse group of curious, passionate people and help us solve a spectrum of autonomous mobility challenges.