I help organisations design and deploy AI systems that survive real-world constraints.
From embedded hardware to cloud-scale AI deployments
I design for latency, cost, reliability, and failure modes
I help teams decide what not to build - before it gets expensive
I advise and build AI systems for organisations operating in real-world conditions - from architecture decisions to production deployment. All consulting and development work is delivered through Obsidian Reach.
I provide senior technical judgement to founders and leadership teams navigating complex system, AI, and infrastructure decisions - helping them avoid costly mistakes before they happen.
I speak on the realities of deploying AI systems beyond prototypes - covering real-world constraints, system design, and the gap between hype and production.
I contribute as a subject-matter expert to media, podcasts, and publications on AI, systems engineering, and technology leadership - grounding discussion in practical experience rather than speculation.
Designed systems that separate genuine environmental events from noisy radio and network artefacts in IoT deployments.
Architected on-device computer vision systems operating under strict power, latency, and reliability constraints.
Built detection systems for high-risk transactional environments, balancing accuracy, latency, and false positives.