Solutions Architect. 25 years building production systems that can't fail.

Commonwealth Bank — Buy Now Pay Later

Before Feneco existed, its founding engineer served as Lead Engineer on the modernization of the Buy Now Pay Later (BNPL) system at Commonwealth Bank — migrating a rudimentary architecture to a cloud-native, serverless solution on AWS that now processes hundreds of thousands of transactions per day, built on Clean Architecture and Domain-Driven Design. That engagement is where Feneco's approach to legacy modernization took shape: disciplined migration under full production load, with zero tolerance for downtime.

At Feneco

Feneco delivers end-to-end solutions for companies that need to move off legacy systems without stopping the business — serving four clients so far, including:

Liren

A custom-built LMS platform Feneco's team designed and maintains across four markets — France, Luxembourg, Belgium and the UAE (liren.fr, liren.lu, liren.be, liren.ae) — serving 15,000+ registered accounts and nearly 6,000 active users.

Read the case study →

BePolyglot

A language-learning platform built around three training modes — a human tutor, a hybrid human-plus-AI-avatar lesson, and a fully AI avatar teacher — with live conversation practice and one progress record across all three. Currently in development.

Read the case study →

StudyBird Australia

A case and proposal platform for an international education agency, running AI entirely on the client's own infrastructure. Passport details, visa history and medical records never leave a single AWS account in Sydney. Currently in development.

Read the case study →

Feneco's team works directly with CTOs, CEOs and Heads of Engineering who need to modernize without betting the whole operation on a full rebuild — disciplined automation and migration, built on top of what the business already trusts.

Private AI

Some data can't leave the building. Identity documents, health records, legal files, financial detail — sending them to a public model API is either a compliance problem or a commercial one. Feneco builds systems that run open-weight models on infrastructure the client owns, in the jurisdiction the data has to stay in, with the capability you'd expect from a hosted service.

That also means knowing where AI doesn't belong. On the StudyBird build, the brief asked for an assistant that would hold course fees and scholarship terms. Language models don't store facts reliably — asked for a figure they don't have, they produce a convincing one, and a quote with an invented fee is a real liability. The design puts a structured database in front: the model retrieves and writes, it never generates a number. A missing fee shows as blank with a warning, not as an estimate.

In practice that looks like open-weight models on your own infrastructure and in your own region, human approval modelled as a state rather than promised in a disclaimer, an audit trail on every AI-assisted decision, and phased delivery where the first stages ship value with no AI at all.

Credentials

• — MBA in AI Engineering, Fullcycle (in progress)
• — Executive MBA in Solutions Architecture, FIAP
• — Post-Degree in Software Engineering, UNA
• — Bachelor's in Systems Analysis and Development, UNA
• AWS Certified
• Azure Certified