I build production software with AI in the loop - turning manual, paper-based operations into systems teams run on.

Cape Town, South Africa

Director & Lead Software Engineer - Technology × Finance × AI Zascia Hugo

( 01 ) - About

Portrait of Zascia Hugo — Portrait - Cape Town

I lead engineering at Webtra Digital Solutions, where I own how software gets designed, built and shipped - work that sits where engineering meets finance.

I take manual, paper-based operations - the spreadsheets, the printed forms, the process that lives in one person’s head - and turn them into systems a business can operate on. Director & Lead Software Engineer at Webtra (Pty) Ltd. A finance background grounds the engineering: I design for reconciliation, audit trails, and controls - not only for clean architecture.

The software runs where the work happens: web applications in the browser, mobile apps for Android and iOS, and desktop software for Windows, Linux, and macOS.

I own problems end to end - mapping the manual process, isolating the decision that matters, and shipping the software that replaces it. On 16Personalities I test as ENTJ-A (Assertive Commander).

I travel to experience new cultures. Cultural exchange is a source of perspective - and of personal growth - that I carry into how I work with distributed teams and unfamiliar domains.

( 02 ) - Tools & Languages

A practical stack - chosen so the systems I ship stay operable years after the first deploy.

The tools below are not a catalogue of everything I have touched. They are the ones that repeatedly earn their place when the work has to leave a laptop and run a business - whether that means a web app, a mobile build for Android or iOS, or desktop software for Windows, Linux, or macOS. TypeScript and Vue on the frontend, PHP with Laravel or Slim when the domain sits closer to a server-rendered API, and relational data in MySQL or PostgreSQL when the records matter enough to query, reconcile and audit.

Infrastructure follows the same rule. Docker keeps environments honest between my machine and production. AWS and DigitalOcean host what needs to stay up. DynamoDB appears where a document store fits the access pattern better than a table. GitHub holds the history; Postman and DBeaver are how I prove the edges before anyone else depends on them.

Familiarity is a feature here, not a compromise. I pick tools a future engineer - including a future version of me - can open, understand and extend without a scavenger hunt through last year’s novelty.

TypeScript

Vue.js

Laravel

PostgreSQL

AWS

Docker

GitHub

Postman

VS Code

Full breakdown

( 03 ) - How I work with AI

AI is integrated into how I engineer - not an add-on. It is how the work gets done.

Treating it as infrastructure changed what I consider a reasonable amount of software for one engineer to own. In practice that means Cursor and Claude as daily drivers rather than autocomplete: delegating breadth - scaffolding, migrations, test coverage, refactors that span dozens of files - while keeping the judgement calls. What to build, how the data is modelled, and which trade-off a system should make remain mine to answer.

The discipline it demands is review. Generated code meets the same bar as anything else that reaches production: read it, understand it, prove it under test. The bottleneck moves from typing to specification and verification - where an engineer’s experience is best spent.