E-commerce architect
Designing commerce platforms that survive their next replatform, and planning the route off the current one.
I am an e-commerce architect with fourteen years on enterprise commerce platforms — Oracle Commerce ATG, Oracle Endeca and commercetools — across B2C retail, B2B distribution and telecom. The work is usually the same shape: a platform built for a business that has since changed, and a migration path the business can actually afford.
The platforms
Deep Oracle Commerce ATG: order management, cart, checkout, promotions, custom pricing engines at item and order level, and the Endeca search layer beside it. That is thirteen years of it, which is long enough to know which parts of the platform are load-bearing and which are folklore.
On the composable side, commercetools — certified as both Associate Java Developer and Associate Functional Architect, though both certificates lapsed in 2022 and I list them with issue dates rather than pretending otherwise.
Replatforming without a freeze
The common failure is a two-year migration during which nothing else ships. The business loses patience, the programme is cut, and now there are two architectures to maintain instead of one.
The approach that has worked: decompose by rate of change, run new services alongside the existing core through contracts you control, and make every increment independently valuable. At Verizon and Rogers that meant pulling an Oracle Commerce monolith into reactive services on Play/Akka. At Kohler it meant Micronaut services on Kubernetes running beside the ATG core.
Integration is where it actually breaks
Commerce is never one system — PIM, OMS, ERP, pricing, payments — and most outages happen at a boundary rather than inside a component. I treat every integration as a contract with an owner and a failure mode, which is less exciting than the replatform and matters more.