CG

From business idea to shipped reality.

AI systems, Apple platforms, cloud architecture and IBM i (AS/400) modernisation, led end to end. I work fluently in both worlds, the boardroom and the codebase, and these are the capabilities I bring to turning a vision into a working product.

Lead pillar

Idea → Reality

Turning ambiguous business vision into modern, innovative digital products that actually ship.

Discovery & problem framing

Cut through a vague brief to the real problem, the user, and the outcome worth building.

Vision → solution architecture

Translate business goals into a pragmatic technical plan and the smallest thing that proves value.

Innovation & R&D

Evaluate emerging tech (AI, on-device, cloud-native) and turn it into a defensible product edge.

Product strategy

Sequence a roadmap that balances quick wins, technical foundation, and long-term direction.

How it ships

Leadership & Delivery

Building high-trust teams and reliably hitting milestones, while keeping directors and engineers aligned.

Team scaling & org design

Grow teams and structures that stay fast and accountable as headcount and scope increase.

Delivery & milestone ownership

Own the commitment end-to-end; de-risk early, communicate honestly, and land the date.

Director ↔ engineer translation

Turn business intent into buildable scope, and engineering reality into board-ready options.

Mentorship & capability building

Coach engineers and leads; hire well and raise the bar so the team outlasts any one project.

Where the work is going

AI Systems

Building AI that survives contact with real users, not demos that only work on the happy path.

AI agents & LLM integration

AI agent development and LLM integration, self-hosted or managed.

RAG & knowledge systems

Custom retrieval over private knowledge bases with pgvector, Pinecone, and Bedrock inference profiles, plus the semantic caching and memory management that keep it affordable.

Real-time voice agents

WebRTC voice agents with separate STT, LLM, and TTS models, tuned for low-latency turn-taking.

On-device AI

Apple MLX and Core ML on M-series silicon, and the on-device versus cloud trade-off.

Read the series

Data residency & sovereignty

AI processing kept inside Australian region data centres, across the whole pipeline: the language model, speech-to-text, text-to-speech and embeddings each see the data.

Cost-capped AI assistants

Website assistants on Next.js and Vercel with no chatbot vendor: AI Gateway holding the model key, prompt caching instead of retrieval, and a hard monthly budget cap enforced before the model is called.

Read the build

The depth behind it

Technical Breadth

Genuinely hands-on across the stack. I lead from a place of real technical understanding.

Cloud & DevOps

AWS with Terraform, containerised deployment, CI/CD, and choosing the option that fits the budget and the real traffic.

Mobile & Apple

iOS since 2013, from Objective-C, Storyboards and XIBs to Swift and SwiftUI, including enterprise distribution and App Store publishing.

IBM i (AS/400) modernisation

RPGLE SOAP services migrated to documented REST APIs, with DB2 kept as the source of truth and Infor M3 pricing returning the same numbers.

Multi-tenant & dedicated SaaS

Both shapes: one deployment serving everyone, or a dedicated instance per customer, with tenant isolation and upgrades that take nobody down.

Enterprise systems

Java & Spring Boot, .NET, and the integration layers that hold enterprise systems together.

Architecture

Modular, maintainable systems sized to the team and the load, not the hype.