I take enterprise AI from strategy to governed, production-grade reality.
Head of AI candidate. I lead du's AI initiatives from architecture and governance — setting the use-case portfolio, authoring the AI governance and identity frameworks, and shipping production AI agents.
Hover a node. Most AI leaders own one of these. I connect all four.
A Head of AI has to hold the whole stack — from the board-level call on what to build, down to the data plumbing that makes it run. Here's where I work.
On du's AI architecture board I help decide which AI use cases the enterprise prioritises and how they're delivered — turning ambition into a sequenced, fundable portfolio.
I authored du's AI governance framework and AI identity framework: the controls that make AI responsible, secure and auditable — and that regulated markets now demand.
I design and ship production AI agents, and prototype across LLM, RAG and multi-agent systems — staying hands-on so direction is grounded in what actually works.
I built du's Next Gen Integration Layer and real-time streaming backbone — the governed data foundation that makes AI/ML, lineage and compliance possible at national-operator scale.
The hardest part of enterprise AI isn't the model — it's knowing where it pays off and how each sector's rules constrain it. Nineteen years across regulated, mission-critical environments taught me that. Pick an industry.
My deepest and current domain. At du I lead enterprise AI from architecture and governance, and built the Next Gen Integration Layer and real-time streaming backbone that powers AI/ML, lineage and compliance at national-operator scale. Earlier I ran production middleware environments, capacity planning and enterprise release migrations for Telstra, and trained engineers as part of a middleware centre of excellence.
Authored du's AI governance and AI identity frameworks; sit on the AI architecture board; contribute to TM Forum AI-native architecture and Open API standards shaping how the whole industry builds AI.
Designed and delivered a centralised enterprise service bus integrating nine APAC systems for Citibank in Singapore, accelerating cross-region digital transformation inside a heavily regulated financial environment. The work demanded the kind of resilience, auditability and stakeholder discipline that AI in banking now requires.
I've mapped a practical AI agenda for banking: agentic backend operations for loan processing, KYC and document validation; GenAI that turns complex credit-card statements into clear, personalised insight; and privacy-safe data monetisation as a new value stream.
Delivered integration architecture for Allstate within high-availability, regulated financial-services environments — aligning delivery with strict compliance, resilience and continuity requirements across complex legacy estates.
The same governed, auditable approach insurers require for AI: model and data integrity, lineage, and identity controls — backed by a blockchain-anchored anomaly-detection design that guards AI pipelines against data poisoning.
Led a 20-member hybrid platform team across SnapLogic, TIBCO, Solace, Splunk, Elastic and OpenShift, delivering 80% efficiency gains through DevOps and automation. Architected fault-tolerant disaster-recovery and backup frameworks with strict RPO/RTO targets, and ran capacity and roadmap planning across 24+ hybrid-cloud platforms.
Mission-critical reliability discipline — the operational backbone AI needs to run safely in a business that never stops. Knowing how 24/7 platforms fail is what keeps AI deployments dependable.
Worked across retail data and business-intelligence reporting, and designed privacy-first data products for retail and FMCG use — combining customer insight with strict boundaries on personal data.
Anonymised, aggregated insight platforms that let retailers and FMCGs understand trends, segments and purchasing behaviour — without ever exposing individual customers. AI value with privacy designed in.
Strategy is cheap without delivery. Each item below is labelled honestly — what's running in production, what's being built, and what's a proven prototype.
An AI agent that generates solution architecture diagrams automatically — turning documents and intent into clear, consistent design artifacts.
Validates designs against TM Forum Open API standards automatically, closing a manual, error-prone governance gap at scale.
An agentic system to manage software delivery lifecycles and demand management across the enterprise — autonomous, orchestrated backend operations.
An NLP + vector engine that extracts sentiment, themes and CX trends from Google Reviews and other sources.
An LLM + RAG tool that produces API specs, UML and sequence diagrams from documents — boosting architecture productivity and governance speed.
A blockchain-anchored anomaly-detection model that protects AI models and data pipelines from poisoning and manipulation.
From production engineer to enterprise AI leader — every role added a layer: reliability, regulated delivery, large-team leadership, and now AI strategy and governance.
Lead du's enterprise AI from architecture and governance; AI architecture board member; authored AI governance and identity frameworks; shipped production AI agents; built the NGIL and streaming data foundation; AI RFP and vendor-selection reviewer.
Led a 20-member hybrid platform team; 80% efficiency gains via DevOps and automation; fault-tolerant DR/backup with strict RPO/RTO across 24+ hybrid-cloud platforms.
Designed and delivered a centralised ESB integrating nine APAC systems for Citibank in Singapore, accelerating transformation across regulated financial environments.
Led production architecture, capacity planning and enterprise release migrations for Telstra; TIBCO trainer for Accenture's middleware centre of excellence.
Supported production for Toyota Financial Services and contributed to Nielsen's CTO Office emerging-technology research. Awarded Best Performer.
I'm open to enterprise AI leadership mandates across the UAE, Saudi Arabia, Singapore and Switzerland. If you're building an AI function that has to be real, governed and shipped — let's talk.