Real results from data platforms, enterprise applications and AI systems built for regulated and high-volume environments: what was in the way, what we built, and what changed.
Replaced a fragmented legacy estate with a metadata-driven Azure lakehouse where new sources onboard by configuration, failed jobs recover themselves, and governance held through the first audit cycle.
A governed calculation engine that took a critical month-end process off spreadsheets, replacing it with reproducible results, a full audit trail, and finance running it without engineering in the loop.
Audited and rebuilt an unmaintainable reporting estate around unified definitions and a centralised logic layer, so the same metric now means the same thing in every department.
A fully serverless RAG agent platform for advertising data, built for instant scale during bursts, nothing running when idle, and live prompt and feed control for non-technical operators.
A WhatsApp chatbot that answers students the moment they ask, resolving routine queries from the company’s question bank so the support team is no longer the bottleneck on response time.