Building reusable foundations

Context

I moved to XTRAC, Fidelity's workflow business, first as a senior engineer on the labs team and then as principal architect. The labs remit was explicit: run 8 to 12 week proofs of concept and find the investments that would carry the business forward. Roughly one in five got funded.

The constraint

The business wanted to digitise paper-based services, but every project was building its own plumbing from scratch. Integration, forms handling, environments and releases were re-solved on every project, so each digitisation effort cost as much as the last one.

The intervention

I designed and built XTRAC Connect, a Java integration engine running natively on AWS, as a building block for the digitisation strategy, and defined a forms continuum on top of it so paper forms could be converted into digital journeys repeatably. I automated environment creation with CloudFormation, Ruby and Jenkins until a single person could operate the platform. As principal architect I re-architected the Fidsafe file-sharing platform for a thousand-fold scale increase and led the migration of a Java/Spring monolith to serverless Node.js services on API Gateway, Lambda and SQS.

Outcomes

• XTRAC Connect became the foundation for the digitisation-of-services strategy.
• A 28-page pensions PDF became a digital journey, cutting processing from over a week to under a day.
• Fidsafe was redesigned to serve 10 million+ users cost-effectively, with CD pipelines that took releases from monthly to daily and cut over $400k in the first year.
• Environment provisioning was fully automated; one engineer could run the platform.

Capability gained

Platform thinking. The unit of work stopped being the application and became the capability underneath it: the integration engine, the pipeline, the reusable infrastructure. And cloud-native architecture as a first language rather than a migration target.