Hackathons for Accelerated Vehicle Software Development at Stellantis

The multi-national automotive manufacturer Stellantis (NYSE: STLA) has selected Amazon Web Services (AWS) as its preferred cloud partner for vehicle platforms. Together, the two companies successfully leveraged Hackathons as a means of commencing new projects and achieving continuous delivery for accelerated vehicle software development. This blog post shares both insights on how to conduct Hackathons and the best practices for running them. Specific examples of successful Hackathons conducted with Stellantis are provided within the context of a multi-year project aimed at developing a cloud-native development environment for Software-Defined Vehicles (SDV).

Embarking on large projects can be intimidating, particularly when working on new topics that will transform the way of working within a company. Large organizations with complex infrastructure and mature organizational cultures, like Stellantis, encounter difficulties during the process of adopting agile, which is state-of-the-art in software development to deliver increments and focusing on continuous releases. This includes challenges such as a lack of experience with agile methodologies, lack of management support, and alignment of project team efforts with project goals. In fact, AWS Professional Services (AWS ProServe)’s Experience Based Accelerator’s (EBA) finding indicate that a majority of friction points encountered in such projects are non-technical in nature.

What are Hackathons?

Hackathons are a delivery methodology that revolves around a week of coding on dedicated use cases. The approach focuses on bringing Minimum Viable Products (MVPs) to life through small delivery increments, which are a concrete stepping stone toward the product goal and a first usable version of a product. A key advantage of Hackathons is the absence of hierarchical structures – it empowers the participants to have more responsibility and autonomy and enables an effective collaboration across various functional teams.

Figure 1- The Four Hackathon Phases

Phase 1 – Identification

The first step towards planning a Hackathon is to identify the need for it. Common indicators are:

• The start of a new team on a new topic
• Challenges implementing MVPs or increments
• Teams struggle following an agile way of working due to missing experience

After the initial need is identified, different stakeholders, which are interested or affected by the hackathon’s outcome, have to be involved. These are typically sponsors, industry experts, business owner, potential users or customers.

Phase 2 – Preparation

The key to a successful Hackathon lies in the preparation, adopting the Amazonian mechanism of Working Backwards from the desired customer outcome. Following the mechanism, you specify problems for a specific customer and describe the future customer experience as a use case. This phase involves collaboration with subject matter experts (SMEs) to define clear and measurable goals for each use case, along with corresponding MVPs. By forming cross-functionals teams and assigning them to a specific use case, you ensure a diverse skillset and expertise within each team. It allows teams to work self-sufficiently and resolve blocking issues or dependencies without further support. It’s important to include the team in defining the use cases so they feel an increased ownership of the outcome.

Schedule the event and communicate goals and use cases at least 4 weeks prior the Hackathon to ensure the participants are committed to the event. Figure 2 is showing an exemplary agenda with recommended topics to plan for each day:

• Kick-Off: Invite executives to the meeting and present goals, working mode and organization of the Hackathon to motivate participants and get their support to remove roadblocks. Set clear rules and guidelines for the teams, such as workspace arrangements and communication rules.
• Use case work: The teams work on their use cases during this time.
• End-of-day Check-In: Set up a common meeting with all teams to share their progress incrementally and iteratively. This enhances transparency to all participants and supports the collaboration across the teams.
• Demo: Invite executives and all stakeholder to share the achievements of the Hackathon.

Once the preparation of the event is complete, the groundwork is laid for the exciting and innovative Hackathon to unfold.

Figure 2 – Exemplary agenda

Phase 3 – Hackathon Week

At the heart of the Hackathon is a one-week onsite event that serves multiple purposes. The use case teams build their MVPs towards the defined goal. Stakeholders are able to support use case teams by validating architectural assumptions and swiftly eliminating obstacles to accelerate the delivery of MVPs. Senior executives have unique opportunities to actively engage with their teams, stepping in to remove blockers and leveraging the presence of domain experts to lay out the strategic plans and overarching vision. The demo at the end of the week provides a stage for developers to showcase their functioning MVPs and to celebrate their success.

Phase 4 – Debrief and next steps

Follow agile best practices and organize a retrospective with each team to identify improvement options for the next Hackathon. Furthermore, each team should take ownership in documenting the next steps or required implementation tasks after the event.

Examples of a successful Hackathons by AWS & Stellantis

Example #1: Hackathons as Accelerators:
In March 2023, 70 software experts from Stellantis, AWS and other suppliers came together for a Hackathon in Rüsselsheim, Germany. The goal was to work on use cases for Stellantis’ cloud-based product development environments, called Virtual Engineering Workbenches (VEW). These allow developers to develop automotive software in a virtual environment that fosters collaboration and increases productivity. Read the official press release how Stellantis and AWS work on virtual platform for continuous software development for automotive.

“Our AWS Virtual Engineering Workbench will enable all our developers and partner teams to provision within minutes a complete workspace including tooling and virtualized hardware and just start building automotive software for our customers,” said Heiko Schilling, SVP Global Software Engineering Stellantis SWX.

The hackathon’s use cases included, among others:

1. Supporting the automated building and testing of virtual ECU (vECU) components built on the Virtual Engineering Workbench (VEW).
2. Creating a co-simulation of vehicle functions in software-in-the-loop (SiL) environments on AWS with hardware-in-the-loop (HiL) in Stellantis’ on-premises datacenter.
3. Providing Blackberry QNX based virtual targets to support Cockpit / In-Vehicle Infotainment (IVI) application development.
4. Integrating NICE DCV as a protocol to connect to virtual development environments for an improved user experience.

Use case 1 – Automated building and testing of vECU components: Software developers used pre-configured VEWs to build vECU binaries. The VEW integrates with Stellantis CI/CD pipeline, which automates the software delivery process, to automatically test the developed vECU. This results in validated vECU artifacts and test coverage reports, illustrating the automated integration of vECU development and functional testing. This provides immediate feedback to software developers and testers and a consistent quality in software builds for Stellantis.

Use case 2 – Co-simulation of vehicle functions in SiL environments on AWS with HiL on-premises: Software developers used VEW to develop a function for folding/unfolding the external vehicle mirrors. The function was then simulated on a SiL-environment on VEW and connected with a HiL-device at Stellantis’ on-premise datacenter. In this use case, the VEW provisions and configures both, SiL and HiL, to co-simulate the mirror fold function. By this, Stellantis has more flexibility in software integration and enables a faster software validation through VEW.

Use case 3 – Blackberry QNX based virtual targets: The team successfully created virtual targets for QNX using AWS Elastic Compute Cloud, enabling infotainment and Cockpit use case developers to build and test their applications in the cloud, removing the dependency on real hardware.

Use case 4 – NICE DCV is a high-performance remote display protocol that provides near real-time responsiveness for your application without compromising on the accuracy of the image. For the VEW it has greatly outperformed the previous solution based on RDP.

The above use cases are of high importance as they successfully demonstrate the integration of different phases of Stellantis’ Automotive software development lifecycle (build, test, integrate, validate). They also prove the powerful flexibility of the cloud-based VEW approach of letting the users combine real hardware with virtual devices of varying levels of abstraction.

Example #2: Hackathons to kick-off a project:
Stellantis and AWS leveraged a joint Hackathon just three weeks after the project start. The event aimed to align the team on a shared vision as well as to design and validate core architectural concepts with stakeholders to build a first MVP of the platform. By this project teams and stakeholders should gain trust in the ambitious roadmap.

Within one week, the hackathon teams created a user-friendly self-service portal for VEW, enabling effortless browsing, selection, and deployment of pre-configured runtime environments in dedicated project AWS accounts. Additionally, a first toolchain, used by vECU developers was implemented. The AUTOSAR runtime environments were registered in the platform’s product portfolio powered by AWS Service Catalog and automatically distributed to the target AWS accounts. The multi-account structure and connectivity to critical on-premises systems like licensing servers, GitHub and JFrog Artifactory was enabled by a dedicated hackathon team. The Hackathon’s collaborative spirit, involving experts from Stellantis, AWS, and other suppliers, further reinforced trust in the shared vision by delivering the first MVP of VEW in a collaborative manner within one week.

As an outcome, the teams deployed the first workbench, a pre-configured development environment, from a self-service portal and enabled Stellantis’ software developers to start their job in minutes. The achievements were presented to a group of stakeholders with positive feedback on the live demo, which made the vision more tangible and created excitement for the roadmap ahead.

Conclusion

Hackathons have emerged as a crucial differentiator for Stellantis’ and AWS ProServe’s VEW project. We are proud to count over 600 demo participants and a total of 18 developed use cases across 4 events as of today. Hackathons are versatile tools and can be used not only to accelerate the project start but also to crack the most complicated use cases, reconfirm agile ways of working and mobilize teams that face various challenges. Stellantis and AWS Proserve have made it a project habit, holding multiple sessions a year to accelerate key use cases that require close collaboration.

Follow our best practices and use Hackathons as a mechanism to accelerate your project delivery.