How Oracle Red Bull Racing’s chief engineer fuels racing strategy with data

April 1, 2022 | 7 minute read
Natalie Gagliordi
Senior Writer
Text Size 100%:

Oracle Red Bull Racing warmup lap

An aerodynamicist by trade, Guillaume Cattelani has taken his PhD in fluid dynamics and turbulence modeling and made it his life’s work to solve one very specific problem: Making race cars go faster.

It’s a skill set that comes in handy as part of the Oracle Red Bull Racing Formula One (F1) team and Red Bull Advanced Technologies, where Cattelani is the chief engineer of technology and analysis tools. In this work, Cattelani is leading efforts to connect data and data science to all the mechanical engineering processes that make the heart-pounding racing we see on the track possible.

“In our very specialized niche industry of Formula One, there is a huge amount to be gained by embracing the data science revolution,” Cattelani says. “So, the idea is to provide a bridge between this world, which is highly technical and is moving very fast, and the engineering problem we have, to design a race car. How can we connect those two things to make a better car, to make a faster car, and make a winning car?”

Guillame Cattelani, Chief Engineer, Oracle Red Bull Racing and Red Bull Advanced Technologies
“Becoming a master at modeling and execution in real time is an advantage,” says Guillaume Cattelani, chief engineer of technology and analysis tools for Oracle Red Bull Racing and Red Bull Advanced Technologies.

Technology and innovation are engrained in the ethos of F1, a sport where race cars are precision engineered for speed and performance. For each race, engineers make thousands of refinements to the car, adapting it to the nuances of the different tracks and to conditions as specific as altitude and humidity. The smallest tweaks to a car’s chassis and aerodynamics can shave milliseconds off a driver’s race time and hand a team the advantage. “The car will probably never run in the same configuration twice,” says Christian Horner, the team principal and CEO. “It’s just constantly evolving.”

Increasingly, F1 teams are looking to give comprehensive data analytics to engineers and drivers. With about 150 sensors on the car collecting data, each race produces about 400 GB of data. Oracle Red Bull Racing uses Oracle Cloud Infrastructure (OCI) to accelerate the team’s use of data science and analytics and separately to guide the engineering work performed at the Red Bull Technology Campus. A key focus for Cattelani’s team is the creation of a technology framework and specific tools—what he calls “assistant engineers” (AE)—that will serve as a new connective thread between traditional engineering disciplines and data science.

“Obviously, there was already data science in motion at Red Bull, but we want to generalize it to all other engineering disciplines, because the potential is huge,” says Cattelani. “Every single piece of data produced should be weaponized and should be used for performance, and that requires different mindsets, different technologies, and a different vision.”

AI for engineers

The 2022 F1 season is putting even more pressure than usual on engineering and data analytics teams. This year brings a dramatically different car design and new regulations set by the Fédération Internationale de l’Automobile (FIA). The 2022 car puts an emphasis on the aerodynamic phenomenon known as ground effect. Ground effect creates downforce, a measure of how much vertical aerodynamic load is created by an F1 car’s surfaces. The bottom line for fans: The new design and rules about ground effect and downforce are meant to enable closer racing and the potential for more overtakes, leading to more exciting race action.

“It’s the biggest rule change that we’ve seen in the last 30 years,” CEO Horner says. “So, that means that nearly every single component on the car is brand new this year. And with it being a ground effect car, with it being designed to make overtaking easier, that’s changed the whole philosophy of how we design these cars. It’s a steep learning curve, and it’s a race of development between the first race and the last race.”

At the same time, however, F1 teams face stringent spending limits, so they must be ultra-efficient with the staff time and resources spent on designing and optimizing their cars. To work within that framework, Cattelani says that his team is creating a set of tools, running on OCI, that will use data science and software engineering to help engineers do more with less. Dubbing this technology toolset AE, Cattelani is optimistic about its potential to present engineers with more data-driven insight into design and performance, without having to hire a crew of data scientists.

“Every single piece of data produced should be weaponized and should be used for performance, and that requires different mindsets, different technologies, and a different vision.”
—Guillaume Cattelani, Chief Engineer, Oracle Red Bull Racing and Red Bull Advanced Technologies

“What we would like to do is to allow our engineers to do more complicated things that traditionally were gained by employing more workforce, training more people,” he says. “The idea is to be able to offer our engineers greater capability to understand the car and how they design it, and to understand where to find performance. It’s really knowing the current performance in greater depth, but also trying to leverage more capabilities for a single engineer. This is really in tune with the resources restriction. I think this is the future for us.”

Simulations and strategy win the race

In addition to the work of refining the car, race strategy is another data-driven initiative that Cattelani’s team supports. For both prerace strategy preparation and in-race decisions, the team runs Monte Carlo simulations to generate a range of potential outcomes and the probabilities for each. These simulations help race strategists guide decisions that the driver and crew must make, like when during a race to make a pit stop to change tires.

Oracle Red Bull Racing ran billions of simulations on OCI during the 2021 season, taking advantage of OCI’s on-demand compute capabilities during race weekends. These simulations helped the team design the optimal strategy prior to each race and inform strategy calls during the race.

Ultimately, the human being behind the wheel and in the pit lane makes the decisions during a race, but that decision is informed by more accurate data, Cattelani explains. Horner has said that OCI-powered simulations assisted with race-day decisions that helped Max Verstappen win the 2021 Drivers’ Championship.

“Becoming a master at modeling and execution in real time is an advantage,” says Cattelani. “The more complicated simulations you can do, the better understanding you have on your strategic options. So, that capability is quite important and has proved to be successful.”

Pit Crew works on Oracle Red Bull Racing car

Driver Sergio Perez pulls the Oracle Red Bull Racing car into the Pitlane during qualifying ahead of the F1 Grand Prix of Bahrain, on March 19, 2022.

Most of the team’s data modeling systems run live on OCI during a race, processing a variety of data generated at rapid speed. Looking ahead to the 2022 season, the team intends to use OCI’s flexible architecture to not only increase the complexity and accuracy of the race simulations, but also to expand the simulations to include the potential strategies of its competitors on the track, enabling Oracle Red Bull Racing to make more informed decisions on strategy.

“Compared to some data-dependent organizations and businesses, we are not the biggest consumer of data. But the variety of data we consume is the complexity of Formula One,” says Cattelani. “It’s really about a variety of data and how you can interpret it. So that’s what we’re going to do for the 2022 season: We’re going to increase the dimensionality of our race strategy models to get this extra information we didn’t have access to before.”

Racing into the future

In a project separate from the Oracle Red Bull Racing F1 team, Oracle is working with Red Bull Advanced Technologies on how AI and machine learning might be used to analyze video and digital inputs, providing young drivers with video overlays that compare their lap performance against modeled optimal laps.

Pulling in data streams from video games, carting, F2 and F3 races, Red Bull Advanced Technologies and Oracle are aiming to familiarize younger drivers with the driving style necessary to extract maximum performance from an F1 race car.

The training models are still in development and not in production yet. But the potential value of these kinds of simulated environments is high because, unlike most sports, young drivers have a limited amount of track time with the actual cars, so they rely heavily on virtual environments to improve their driving skills at the wheel of an F1 car.

“It’s quite interesting to understand how a driver interacts with a synthetic environment and how he can potentially interact with a real machine,” says Cattelani. “We want to try to understand in the future, a bit better, the relationship between driver and machine.”

Photography: Courtesy of Red Bull Racing, Mark Thompson/Getty Images

Natalie Gagliordi

Senior Writer

Natalie Gagliordi is a senior writer at Oracle. She spent a decade as a technology journalist and was previously a senior writer on ZDNet.


Previous Post

Innovation in increments for Heathrow using Capgemini’s Agile Innovation Platform supported by Flexagon’s FlexDeploy

Andy Bell | 6 min read

Next Post


Announcing support for network security groups and larger payloads in API Gateway

Robert Wunderlich | 2 min read