Ryan Susana

Air France-KLM.As technical architect for the middleware team of about 150 people, I architected solutions that more than 3,000 applications depend on¹.1Note 1: 3,000+critical and non-critical applications, moving many gigabytes of data a day. I also delivered six enterprise Terraform modules used across 1,000+ applications and integrations.

NPO.In a Google Cloud data engineering team handling 5K–50K events a second, I helped cut the team’s cloud costs by 70%².2Note 2: −70%tens of thousands of euros a month, so far. I also stabilised a pipeline that had weekly outages, including turning it from stateful into stateless.

DICTU.Brought in as solution architect for RDI, part of the Dutch government, to prepare its applications for modernisation across Azure, on-premise and GovCloud, I started with a fit-gap analysis of dozens of applications³.3Note 3: against Dutch law, regulations and industry standards; the work covers BIO, DigiD, identity and data classification.

FedEx.As solution architect for the Quote Platform, a platform responsible for hundreds of millions in revenue a year, I designed its move to Azure Kubernetes Service, worked on it hands-on and guided the team of about 30 developers through it. I also architected splitting a database shared by 23 microservices⁴.4Note 4: 23microservices moving to a database each; partly completed during the engagement.

Bol.com.As a senior software and data engineer, I was part of the team turning raw user events into search journeys and recommendations, about 5 TB of them a day⁵.5Note 5: ~5 TBraw user events a day, from millions of users. I also introduced the team to mutation analysis to measure test quality, and mentored engineers.

ReBound.For a critical payment service, I helped bring integrations with all major Dutch payment providers live within two months⁶.6Note 6: 2 monthsfrom start to live, under PCI-DSS.

The work I like best starts with a tangled system: a database shared by 23 services, a pipeline that falls over every week, a cloud bill far higher than it needs to be. I help take it apart until the better design is obvious, and I can usually show why the old one was worse⁷.7Note 7: I spoke about exactly this at JFall: “Identifying bad design, objectively”.

I set a clear direction and give teams the scaffolding to run with it: C4 diagrams, runbooks, a lot of conversations and written-down decisions⁸.8Note 8: I reintroduced Architecture Decision Records to the Air France-KLM middleware team. The people building it can then decide without waiting for me.

And I stay hands-on. Fourteen years in, I still write the code, the Terraform and the pipelines, because an architecture nobody can build is just a drawing.