[01]
AI agents
Agents that read enterprise knowledge and take action: Copilot Studio, Foundry Agent Service and MCP tool integration, with human approval at high stakes steps.
Copilot Studio / Foundry / MCP
Jeffrey Chu
AI-102 · AZ-305 · SC-100
PMP · 12+ years cloud and ICT
AI Solution Architect · Sydney
Agents you can trust.
Data you can govern.
I design and deploy enterprise AI agents, the inferencing that powers them, and the Microsoft Purview governance that makes them safe to switch on.
Foundry
Agents, inference and evaluation@ in-region model deployments
Labels, DLP and DSPM for AI@ before any agent sees the data
Purview
Copilot Studio
Business agents with MCP tools@ human approval on high stakes steps
What I do
From governed data to agents in production.
AI is only as trustworthy as the data it reads and the controls around it. I work across the full stack: data, governance, models and the agents people use.
[02]
AI inferencing
The right model for each job: Microsoft Foundry and Azure OpenAI model deployments, serverless or provisioned capacity, balancing latency, cost and in-region data residency.
Models / Latency / Sovereignty
[03]
AI data
Enterprise data ready for AI: retrieval with Azure AI Search, permission aware grounding and evaluation, so answers are accurate and only show what each user may see.
RAG / AI Search / Evals
[04]
AI governance and Purview
Purview sensitivity labels, DLP and Data Security Posture Management for AI apps and agents, oversharing controls before Copilot rollout, and Responsible AI guardrails.
Purview / DLP / Responsible AI
Foundation: Azure landing zones, Zero Trust and Microsoft Sentinel, from 12+ years of cloud architecture in banking and regulated sectors.
Credentials
Certified in AI, architecture and security.
AI-102Azure AI Engineer AssociateAssociate
AZ-305Azure Solutions Architect ExpertExpert
SC-100Cybersecurity Architect ExpertExpert
AI-900Azure AI FundamentalsFundamentals
PMPProject Management ProfessionalPMI
How I work
Govern the data first. Then let the agents loose.
- 01 / READINESS
Assess use cases, data quality and oversharing risk; rank the AI opportunities worth funding.
- 02 / GOVERN
Classify and protect data with Purview, set AI usage policy and audit before any agent touches it.
- 03 / BUILD
A working agent in weeks, grounded on governed data, with evaluations that prove it answers correctly.
- 04 / RUN AND SCALE
Monitor quality, inference cost and risk, then extend to the next use case on the same platform.
Jeffrey Chu
Sydney, NSW
Sydney, NSW
Planning an AI agent or Copilot rollout?
Happy to talk use cases, data readiness and governance, for organisations across Australia.