After analyzing the Zoox Careers Page, I came up with 298 different meanings for the position "Self-Driving Car Engineer".
What's a self-driving car engineer? To many, getting one of the many available self-driving car jobs means working in Machine Learning, AI, and Computer Vision. But what's really required for a company to build a self-driving car?
In this article, I'd like us to analyze Amazon's billion dollar self-driving car startup Zoox. This company has nearly 1,500 employees working in 3 different locations, from Foster City, CA to Las Vegas, NV, to San Francisco, CA.
Did you ever wonder: How can a self-driving car company work with 1,500 people at more than 3 different places? What are the different teams? Their relationships? Or maybe "How do I get a job at Zoox?".
In this article, I'm going to explain the Zoox company, but I'm also going to build a framework that will work for any self-driving car company.
The 3 Elements of a Self-Driving Car Team
To understand how the company works at scale, we need to get the 3 core elements:
- Divisions — The high-level topics of the companies like legal, software, hardware, ...
- Teams — The mid-level teams involved in a division, such as perception, localization, marketing, ...
- People — The employees working inside the teams.
Let's play a game 🤡 — There are 1,426 employees working at Zoox, how many divisions do you think there are? 3? 7? 10?
👉 List all the divisions and teams you think are needed in a self-driving car company.
The answer might surprise you, but after 3 hours studying Zoox's Career page, I noted 17 Divisions and 65 Teams!
In this article, I'm going to list the 17 divisions, and explain how they work with each other. We'll also dive into some specific teams such as the Software Team. Each of these divisions is made of teams, full of employees working to make the self-driving car dream possible.
To make it easy to understand, I decided to split the entire company into 7 families, you can click on any of them to directly access the jobs:
Let's begin with the Hardware Divisions.
There are 3 internal Hardware Divisions:
- Advanced Hardware Engineering
- Vehicle Development
- Industrial Design
Advanced Hardware Engineering
The first division I want to take a look at is called Advanced Hardware Engineering.
The goal of this team is to design all the hardware and computer parts. The 8 sub-teams are Digital Signal Processing, Electrical Engineering, Firmware, Hardware Engineering, Hardware Test Engineering, Mechanical Engineering, Sensor Engineering, Systems Engineering & Integration.
What's to notice?
Well, it's mostly engineers. And it's mostly related to the autonomous part — This team doesn't build seats and steering wheels, they build the sensors and calculators that will be linked to the AI code. Something else to notice, even though there's an entire division required to QA and testing, this division has its own "Hardware Test Engineer" team.
The next part is this time building the vehicles! In case you didn't know, here's what a Zoox vehicle looks like:
Now, let's take a look at who's building this. We have 3 teams:
- Program Management
- Vehicle Integration Engineering
- Vehicle Systems Engineering
I won't go too much into details, but these 3 teams are here to build the car. The picture should give you a more precise look.
Finally, the industrial design division is in charge of thinking and designing the future cars of the company. We have 3D, VR, and many "design" jobs.
Something interesting is that these 3 divisions all work in Foster City, CA — so we can assume they work together and communicate to an extent. Next, let's see the operations.
The goal of the operations divisions is to coordinate and organize the entire processes of the company, from daily testing, to fleet management, to how and when the cars are produced. I noted 3 operations divisions:
- Fleet Operations
- Manufacturing Operations
- Program Management Office
In this first part, the Fleet Operations Division is supposed to handle the fleet of cars. The Fleet division is separated into 2 teams:
- Alpha Operations — A team in Las Vegas, NV. I understand it as the Lead team. It's responsible for understanding operations data, ensuring effective communication between development and operations, and informing key decisions to successfully scale the organization.
- Base Operations — Teams in San Francisco and Foster City, CA responsible to keep the documentation up to date, from the software releases to the demos, to organizing everything the company is doing.
Let's zoom into a specific job title: Operations Specialist — First Shift.
Here's the job description: "You will be responsible for making sure vehicle calibrations are up-to-date and statuses are reported across multiple teams. Additionally, you will help to ensure fleet vehicles are primed and ready for daily use by deploying software and troubleshooting both new and existing issues; thereby, directly contributing to the mission success of Zoox.".
Then, when looking at the requirements, you can see this:
So, it's not an "Engineer" job, it's mostly an... operations job!
Next, the manufacturing operations team is responsible for manufacturing the components. They are Quality Engineer, Prototype Fabricators, Technical Program Managers, or Logistics Engineers.
Program Management Office
Finally, the last "Operation" Team is called PMO. According to someone working there, it's "like the music director of an orchestra" — they're responsible to coordinate the company, do people operations, and more.
So far, we've seen 6 of the 17 divisions inside Zoox: they're responsible for hardware and operations. Next, let's see the People Teams.
The goal of the People Divisions is to hire talents, manage them, sell shuttles, or communicate internally. It's 3 main divisions: communications, experience, and people experience.
Every company needs to have a marketing team. Engineers hate them, and they hate engineers... 🙃 But without a marketing team, there is no company!
When you're getting inside a self-driving car, you are living an experience. From the ads you might see on screens, to the comfort, to the feeling you have when looking at the car for the first time. This team is responsible for that:
Finally, the second Experience Team is in charge of the talents — from recruiting to helping them evolve in the company.
As you can guess, the 3 divisions are in the headquarters of Zoox, in Foster City, CA.
The Suits 👔
Next, let's see the suits — These are the finance experts that sold the company to Amazon for 1.2 B$ in June 2020, that scaled the company to 1,426 employees, that handles phone calls, lawsuits, regulations, legal issues, safety, and everything that involves wearing a suit in a room too hot to handle.
No offense, but I'll treat the 4 divisions in a single slide:
There are very few engineers or technical program managers here, it's mostly strategy, legal, or executive jobs.
So far, we've seen divisions for People, Suits, Operations, and Hardware. Now, let's see the second part, we have 3 families to look at: Information Technology, QA, and Software.
Let's begin with IT.
Ever been to a big company? There is always a hot room that smells like heating computers, where you need to go on your first day to get your computer, on your second day to get the stuff they forgot to give you before, and on the day you leave the company. These are the IT people you never work with, but without whom no big company can survive.
The Information Technology division is made of 4 teams: compute infrastructure (VMWare, Windows, ...), Enterprise Applications (all the software you need access to), Enterprise Security, and Networks.
Ever heard some people say "self-driving cars aren't safe! 🤓" Well, the goal of the Quality Assurance divisions is to give these people wrong. The QA team is responsible to test vehicles and algorithms through 2 respective divisions: Corporate Quality and Systems Design and Mission Assurance (SDMA).
The Corporate Quality team is making sure that the cars are safe. They're testing general assembly, EV powertrains systems, body & closure, and more...
You can already see a Data Science job here, the description:
It looks interesting, if you have a Data Science background and don't feel like handling the Perception part yet. In general, I've seen a lot of engineers start with QA, and then transition to coding when they feel ready.
Systems Design and Mission Assurance (SDMA)
To illustrate how they test their algorithms, I'd like to share a video recently published by Zoox where they explain how a test-case is designed and created.
In this video, you see an entire test, cars, and many more created just to test algorithms over and over again. In 2020, I shared an article about the test track I used in Paris when building self-driving cars, you can read the article here.
Here's how the team works:
The 2 test teams are vital to the company, but also to secure investors, politics, insurance companies, and medias...
Finally, let's see my favourite of all: the Software Division. This is what most people see as "self-driving cars". And we've just seen that it's actually completely wrong. Self-Driving Cars is also the 16 other divisions we've just looked at!
But still, without software, no autonomy.
The Map is so big I can't even share it in one image, there are 18 teams just for this alone! I generally split the world of software into 4 pillars:
Here, there are many more!
Let's begin with the parts related to autonomous driving.
Perception, Localization, & Data Science
If you believe that "Self-Driving Car Engineer" means "Perception", you're wrong. Here's the proof with the first 3 teams:
There is an entire team for Data Science, with Data Scientists and Data Engineers; another one for Localization and Mapping, and one for Perception. If you don't know what these terms mean, I invite you to enroll in my course SELF-DRIVING CARS: Break Into the Cutting-Edge World where I detail completely how to join these teams.
The Perception team is made of Machine Learning Engineers, Perception Engineers where I assume the Sensor Fusion talents work, Computer Vision experts, RADAR Engineers, and even Prediction Engineers (it's sometimes seen in the Planning module).
You can take a closer look at the Autonomous Driving algorithms implemented here. Here's a video I liked about Zoox driving in San Francisco.
By the way, if you'd like to become an engineer working on these technologies, I have a platform full of courses on Perception and Localization, you should check it out.
Embedded Software & Systems Integration
Next, these are the 2 teams that handle embedded programming — they make the hardware communicate with the software. They know Linux, ROS, and other integration tools.
Motion Planning, Control, & Simulations
Next, we have the teams that teach the car how to go from A to B, when to overtake a vehicle, or what is the optimal trajectory to follow.
You can see Motion Planning Engineers, Control Engineers, and 15 other job titles. We also have the Simulations Team, responsible for handling scenarios in simulation environments right before they test it in real life. They're basically video game developers, who also know how to integrate the self-driving car software.
You can also see some TeleGuidance engineers. These are people detecting an event, and sending apath to the vehicle. Here's a demo:
As you can see, this is pretty neat!
Rider & Visualization
Next, we have the people who design the apps to order the vehicles. They are iOS and Android Experts, and they handle the customer experience.
Software & Architecture
Then, 3 teams are building software for the company in general. They have a platform to handle the data they collect, they're building maps of the environments, middlewares, mechatronics systems, they create visualization tools, and more...
Cybersecurity, Safety, and PM
Finally, the last set of teams is another QA department inside the software team, responsible to handle vulnerabilities and run tests. You can see Quality Assurance Engineers, Cybersecurity experts, and technical product managers.
Conclusion — Putting it all together
As you can see, the number of jobs is huge! You've just seen what are the positions inside a billion dollar self-driving car company. And indeed, there is a lot to remember!