I make systems answer in milliseconds—
and AI answer correctly.

I grew up in Sydney and cut my teeth on olympiad problems — the kind where a clever bound or the right data structure is the difference between an answer and a timeout. That instinct never left. Today it shows up as retrieval that stays off the latency path, schemas that skip migrations, and tuning loops that finish in a third of the time.

At Voqo I worked across the stack of a live voice agent: the retrieval that grounds it, the evals that keep it honest, and the call orchestration that hands a caller to a human without dropping the line. I care about the boring guarantees — idempotency, precision, reconnection — because that’s what “real-time” actually costs.