I rebuild the network infrastructure behind Australia's largest telecom — and build the machine learning that watches over it.
Load balancers, firewalls, mail relays, web proxies — migrated and re-architected to cloud-native. Then the telemetry pipelines and models that keep the whole estate observable. Deep network engineering and applied ML, at national scale, is a rare combination. It's also the whole point.
Scale
~300 apps
Own the Azure landing zone as product, and carry it through the cybersecurity development process — the same person across all three. In large regulated organisations that gate is usually where delivery dies; I've spent enough time on both sides of it to get builds through. migrated onto an Azure landing zone, under Telstra's five-year strategic partnership with Microsoft — with the governance, IAM and network model that made mass onboarding possible.
Latency
30 hrs → 10 min
re-engineered a security telemetry pipeline from a 30-hour batch to near-real-time, giving an organisation-wide live view of cloud security posture behind one of Australia's largest AI investments.
Cost
~$1M / year
cut from cloud spend by consolidating 400+ ad-hoc subscriptions into a governed hierarchy — then standardising provisioning so the saving held instead of drifting back.
Range
infra ↔ ML
fluent from a Terraform module and a Palo Alto rule to a scikit-learn pipeline.
Leading a business-critical infrastructure modernization at Telstra.
restore momentum on a long-running, business-critical migration, then took the F5 load-balancing tier from lift-and-shift to cloud-native — tripling its resilience.
Notes on cloud architecture, modernisation, and the occasional war story.
Working through ideas in public — landing zones, infrastructure modernisation, security telemetry, and what actually moves the needle inside large regulated organisations.
Read the blog →