The Reliability Mindset : A Culture of Excellence Driving Operational Excellence At Estreem

Picture a nuclear power plant.

Now put it underwater. Cut off all logistics. Cut off all communications. Send it into hostile territory. If the reactor fails, the vessel and its entire crew go down with it.

That is in a nutshell the challenge of a nuclear submarine is.

Out of more than 200 nuclear powered ship built by the US Navy since the 1950s, not a single vessel has been lost due to its nuclear core. Not one. Despite decades of operations in the most extreme conditions imaginable.

This is not luck. This is the result of a culture of excellence. And this very culture applied to critical systems thrilled many other domains like spatial, aeronautics, hospital and now Service Reliability Engineering. Reaching this excellence level is inspiring us at Estreem.

Being performant is not enough

There is a persistent misconception in the tech world: the belief that good processes automatically secure good outcomes. That by auditing everything, documentingeverything, controlling everything, excellence will naturally follow.

NASA in the 1970s was the perfect embodiment of that belief. Half a million people. The finest engineers in the world. Processes audited at every level, phone conferences recorded, total traceability. The ultimate machine-like organisation.

And yet, the machine broke. Twice. Challenger in 1986. Columbia in 2003. Two crews lost. Two public commissions of inquiry made possible by American transparency around the use of taxpayer money. And twice, the same verdict: the problem was neither technical nor procedural. It was poor human decisions, made despite every safeguard in place.

NASA wasn’t aiming to improve performance, it was trying to break through it to reach excellence.

The secret of the most reliable organizations on Earth

After the Columbia disaster, NASA asked a bold question: what is the most reliable organisation in America, and what is the difference with us?

The answer came from the deep: the US Navy’s nuclear engine submarine division, founded byAdmiral Rickover in the 1950s. A smaller organisation. A hundred thousand people versus half a million. And a flawless track record.

What surprised the investigators was that the difference did not lie in processes. Both organisations had rigorous processes. The difference was somewhere else entirely. It was in the culture .

An idea that transcends eras

These principles did not wait for cloud computing to prove their worth. Admiral Rickover went on to lead America’s civilian nuclear power program. That culture, later adopted by Westinghouse, became the foundation of France’s civil nuclear fleet. In 2008, the healthcare sector made a striking discovery: introducing a simple checklist managed by nurses in surgical blocks cut mortality and post-operative complications in half. Even in heavily processed surgeries, uplifting team dynamics saved lives.

Then Google codified Service Reliability Engineering principles. Amazon wrote its Leadership Principles. Look closely, and they are near-perfect transpositions of what the Navy had been practicing since the 1950s. SRE culture may not have originated in the cloud. It was born from a conviction: excellence comes from culture, not from procedures.

What we are building at Estreem

Estreem was born from an ambition: to reinvent payment infrastructure in Europe. Not by replicating existing models, but by building from scratch a SaaS payment platform with a level of rigor unprecedented in this market.

What we want to avoid is operating in reactive mode. Classic production team are at the received end of the delivery, receiving what others have developed and tested, and spending 60 to 80% of their budget for incident management and continuous improvement. Thirty minutes to connect, diagnose, and fix. In the world of SaaS payments, that is an eternity.

We aim ultimately for the five nines standard – 99.999% availability. That means a maximum of 23 seconds of downtime per month. It is extraordinarily demanding. It is also entirely possible – provided your team’s culture matches the ambition.

A partner, not a gatekeeper

At Estreem, the value chain is clear: Product imagines, Software Engineering builds, SRE operates and delivers. As SRE, we are the third pillar. And our first mission, the one etched into our DNA, is to help developers ship faster.

Every business domain has a dedicated SRE counterpart. Their role: be the development team’s closest ally. Talk every day. Understand the stakes. Accelerate delivery. Train developers to improve delivery. Increasing velocity reinforces stability as more frequent updates of leverages incremental changes.

But that partner also carries a critical responsibility: knowing when to pause delivery to address underlying issues. If the code being shipped threatens stability, the deployment can be blocked. More than that: they can halt all ongoing development so the team can focus on resolving the technical issue. It is an enormous power. And it is precisely what makes velocity sustainable.

This mechanism only works in a culture of trust and candor. The very same culture as the submarines: the right to say no, the absence of blame, a shared mission.

Freeing human time

Amazon coined a word that perfectly captures our ambition: autonomation. Autonomy through automation.

Our goal: every service the SRE team provides should be consumable as an API. No one should call us, email us, or wait for our reply. Everything that can be automated must be.

Not to replace humans – but to restore their nobility. If our engineers spend their days processing tickets, they are not spending them thinking, innovating, or engaging in meaningful dialogue. We want every hour of human work to be quality time spent building, improving, and growing together.

See, understand, act

Our team rests on three complementary squads. Domain counterparts, embedded within each business domain, serve as the living bridge between reliability and velocity. The Observability & FinOps squad builds the visibility the entire organization needs – real-time dashboards and data, giving everyone the ability to understand what is happening in our cloud. And the Core Platform ensures the entire technical landscape runs like clockwork, from cloud foundations to production.

These three squads serve one shared objective: enabling Estreem to grow. To onboard new banking partners. To integrate each client rapidly. To innovate on payment APIs without ever compromising stability.

Craftmanship in building to perfect our operations

A nuclear submarine takes ten years to build then thirty years of operations. At Estreem, we share that philosophy. We are in the early stages of construction. We are internalizing skills. We are recruiting people who are not just technically excellent – but who know how to listen, to question, to push back when necessary, and to learn relentlessly.

We are not trying to duplicate classic operations. Nor are we trying to copy Google or Amazon. We draw from the same sources they do – a culture of excellence born inthe 1950s, battle-tested in the most extreme environments on the planet – and we are adapting it to our own reality: European payment infrastructure for the 21st century.

Because a tits core, SRE is not a technical discipline. It is an act of faith in people. It is the belief that the best systems are not built by the best machines, but by the best teams. Teams that dare to pursue excellence – not just performance.

“We don’t want to be performant. We want to be excellent. And excellence is not about process. It’s about culture.”

Mission before title

No delegation to a “safety officer.” Delegating safety to a single person would dilute responsibility. Safety is embedded in every crew member’s mission, without exception.

The right to say no

In a military organisation, this seems counter-intuitive. And yet: every crew member can – and must – challenge an order if it end angers the mission. Even more radical: when a decision is reached unanimously, it is cancelled and resumed with a designated opposite. Unanimity is not a sign of consensus. It is a warning signal.

Never blame, ever

An incident is never one person’s fault. It is a system failure. Mistake punishment eliminates transparency. People may stop speaking up. And the near-misses visibility helps to build safer systems.

Never stop learning

A nuclear submarine takes ten years to build. The crews are hired from day one of construction and participate throughout the entire process. They then spend their whole career aboard the same class vessel. And their very first action once deployed? Deliberately introducing reactor failures to train for real-world repairs. Learning is not a phase. Continuous learning is a mandate.