Agentic AI · Case study
Engineering secure agentic systems for the modern workplace
Technical leadership across AI agents, MCP, applications and Azure platform controls.
Problem
Agentic workplace capabilities need to move beyond demonstrations while protecting organisational data, identities and operational boundaries.
Contribution
I led technical delivery across the Microsoft-focused team, connecting AI agents and MCP servers with Azure AI Foundry, .NET, Aspire and Dapr. I worked across application design, platform security, identity, deployment and governance rather than treating AI as an isolated feature.
Outcome
The programme gained a joined-up architecture and delivery approach for building useful agentic capabilities within production-grade security and platform constraints.
The problem
Agentic AI introduces a new set of interactions between models, tools, enterprise data and user identity. A compelling prototype is not enough: the solution also needs traceable boundaries, safe tool access and a platform that teams can operate.
My contribution
As technical lead, I worked from architecture through hands-on development. I brought together AI agents, MCP servers and cloud-native application patterns with Azure identity, security and governance. Aspire and Dapr supported a coherent developer and runtime experience across the system.
The outcome
The work created a practical route from AI experimentation to controlled delivery. Application behaviour, platform capability and security decisions were considered as one system, helping the team make clearer implementation choices.