At QCon London 2025, Matthew Liste, Head of Infrastructure at American Express, shared invaluable lessons from over two decades of building cloud-based, scalable, secure, and stable infrastructure platforms for mission-critical software in the financial services world. His talk offered a perspective on the delicate balance between constant innovation and the systemic significance of these foundational platforms.
Liste, whose extensive experience includes tenures at JPMorgan Chase and Goldman Sachs, emphasized that building platforms for applications like electronic trading and credit card processing demands attention to uptime and resilience.
Liste began by defining a platform as the bedrock upon which applications and processes are built, drawing an analogy to physical platforms that hide complexity. He stressed that the core role of a successful platform is to abstract underlying intricacies, making it intuitive and easy for consumers (developers and application teams) to use.
A key principle highlighted was the necessity of building standard, interchangeable components. Using Lego bricks as a metaphor, Liste explained how a finite set of standardized building blocks allows for the creation of an infinite number of complex structures. Platforms should dictate standard components for areas like observability and identity management, ensuring seamless integration and a consistent consumer experience.
Liste emphasized the non-negotiable “three S’s ” for platforms supporting mission-critical financial systems: Stability, Security, and Scalability. Unlike real estate, where one might compromise on location, price, or size, platform builders must deliver all three without exception, not just at launch but continuously. He acknowledged the inherent tension between these pillars, such as the need to patch for security potentially impacting stability, and the constant requirement to build for significant future scale.
The concept of being “evergreen” was also crucial. Liste stressed the continuous effort required to maintain and update the underlying environment – potentially comprising millions of components – without causing significant customer disruption. He noted that while cloud-native architectures can ease this, many legacy financial systems require careful planning and execution of upgrades and patching cycles.
Liste cautioned against “undifferentiated heavy lifting,” the temptation for engineering teams to overbuild or reinvent the wheel. He advocated for focusing on clients’ core value propositions and leveraging existing, proven technologies where possible, wrapping them with the necessary controls rather than rewriting them from scratch.
Complementary to this was the principle of being “opinionated.” Platform teams must be willing to say “no” to excessive client demands and focus on delivering the most value to the majority. Trying to accommodate every niche requirement can lead to unsustainable complexity and dilute the platform’s core strengths. Liste shared his experience of learning to say “no” more often to maintain focus and long-term sustainability.
The final principle Liste discussed was to “be long-term greedy.” This emphasizes making decisions based on the platform’s sustainability over the years, rather than short-term gains. Using the analogy of getting a dog – a commitment for a decade or more – he urged platform teams to consider the long-term support and maintenance implications of every new component or feature they introduce.
Beyond technical principles, Liste touched upon the importance of a strong team culture and leveraging open-source communities for shared resources and innovation. Collaboration and empowerment within the team are crucial for building and sustaining great platforms.
Matthew Liste’s talk provided practical insights for anyone building and maintaining platforms at scale. The emphasis on intuitive design, interchangeable components, the non-negotiable three S’s, the need to stay evergreen, focused, opinionated, and long-term greedy offers a valuable framework for navigating the complexities of platform engineering, particularly in demanding industries like financial services. Attendees were encouraged to reflect on these principles and apply them to their own professional practices.
