My work - Proven solutions for real-world problems.

I built and own the data platform behind Decarb.Earth: a single pipeline that ingests telemetry from fifteen inverter ecosystems, processes global solar-irradiance geography in C++, and powers both a public impact map and an internal fleet-health dashboard.

The platform spans 24,350 monitored plants across 44 projects, turning raw inverter readings into audit-ready carbon impact figures.

A self-initiated study built the way I build client reporting: a reproducible pipeline over 10 public sources, committed datasets, and seven interactive charts.

The strongest single relationship in the data is fertiliser cost leading South African food inflation by 10 months (r = 0.52 over 209 months), and the fertiliser basket has just climbed from 120.6 to a peak of 206.0, easing to 162.4 by June 2026.

Tekolo is my own product: offline-first, SARS-compliant invoicing and bookkeeping for South African small businesses, with no monthly fee and no cloud dependency. One Zig core carries all business logic: shipped today as a Tauri desktop app, with a native Linux shell in progress and a WASM-powered web app planned.

It appears here as evidence of engineering standards rather than as client work: money handled as integer cents with banker's rounding at a single boundary, server-authoritative totals, SARS validation enforced at the FFI boundary, and 334 automated tests including no-mock contract tests.

Nestify optimized its short-term rental listings using data science. By creating a data pipeline, utilizing coordinates, and applying clustering, I developed a predictive model for pricing optimization.

The solution increased host revenue and platform competitiveness, establishing Nestify as an industry leader.

In 2017, I created a mobile interface for Hello Group to conduct surveys among underserved rural and migrant communities in South Africa.

The project aimed to test the feasibility of a new banking offering, collecting crucial data on customer segments, purchase intent, product preferences, and desired customer experience.

I developed an NLP solution for Hollard, a leading South African insurer, to analyse customer sentiment across multiple languages. The project resulted in a real-time PowerBI dashboard, automated alerts for negative feedback, and competitor tracking, significantly enhancing Hollard's customer relationship management.