About
AI + Machine Learning + Enterprise Technology + Cloud + Data
I design and ship production AI systems — RAG pipelines, autonomous agents and machine learning services — and integrate them into the enterprise platforms, APIs and cloud infrastructure businesses already run on.
I started in enterprise platform engineering, moved into machine learning, and now focus on generative AI systems. That path matters: most AI projects don't fail on modelling, they fail on integration, governance and evaluation. I've worked on both sides of that line.
Career journey
How I got here
Enterprise foundations
Started in enterprise platform engineering — APIs, integrations and regulated delivery in financial services.
Data and platforms
Moved deeper into customer data platforms, segmentation and large-scale integration work.
Machine learning
Shifted into ML engineering: feature pipelines, model training, calibrated scoring APIs and drift monitoring.
Generative AI
Focused on LLM systems — RAG pipelines, structured outputs, evaluation harnesses and guardrails.
Agentic systems
Building agent workflows with planning, typed tool use and validation gates for enterprise automation.
Engineering philosophy
What I optimise for
Four convictions that shape every architecture decision I make.
Grounding over fluency
A confident wrong answer costs more than an honest abstention. Retrieval quality and citations come before prompt polish.
Measure before rollout
If a system's quality can't be described with a dataset and a metric, it isn't ready to face users.
Boundaries around autonomy
Agents get typed tools, step budgets and validators. Autonomy is a design decision, not a default.
Enterprise reality first
Auth, access scoping, cost, audit and monitoring are part of the build, not a hardening phase afterwards.
Technical strengths
Where I'm strongest
AI & Generative AI
Machine Learning
Data & Engineering
AI Development
Cloud
Enterprise Technology
Key achievements
Professional impact
Grounded by default
Every retrieval system I ship answers with citations and abstains when evidence is missing.
Evaluated before rollout
Golden datasets and automated eval harnesses gate releases instead of subjective spot checks.
Safe automation
Agents operate under step budgets, typed tool schemas and validation gates before any write action.
Enterprise-ready delivery
AI services ship with auth, quotas, redaction, monitoring and cost attribution from day one.
Certifications
Verified credentials
2022
Salesforce Marketing Cloud Developer
Salesforce
2023
Salesforce Data Cloud Consultant
Salesforce
2022
AWS Certified Developer – Associate
Amazon Web Services
2019
ITIL Foundation
AXELOS
2017
Java Certification
Oracle