Dennis Milton
Enterprise Architecture, Migration & AI Systems
Senior Solution Architect · Cloud & Identity Architect · Migration Lead
Copenhagen / Aarhus
I design and deliver technology change in environments that cannot simply be switched off while they are being rebuilt.
For more than 17 years, my work has centred on complex infrastructure, identity, cloud and collaboration platforms — from understanding legacy environments and hidden dependencies to designing the target architecture, migration path and operational handover.
Architecture & Transition
Architecture is only useful if there is a realistic path from the current state to the target state. My work typically covers:
- Enterprise and solution architecture
- Cloud and identity architecture
- Migration and coexistence design
- Legacy dependency analysis
- Technical troubleshooting and recovery
- Security and operational requirements
- Transition into production and operations
I tend to work where architecture and execution meet: systems are already running, dependencies are imperfectly documented, different teams own different parts of the problem, and the business still needs to operate throughout the change.
Migration
Large migrations are not primarily data-moving exercises. Identity, access, applications, mail, files, collaboration, security and operational processes are interconnected. A migration therefore has to work on both sides of the transition — not only after the final user or workload has moved.
The objective is to understand what can break before it does, design coexistence where it is required, and gradually remove the dependencies that made coexistence necessary.
Troubleshooting Complex Systems
A significant part of my work starts when something does not behave as the architecture says it should. That means working across platforms, infrastructure, applications, identity, networking and operational processes to find the actual dependency or failure point rather than treating the visible symptom.
I have worked on everything from enterprise migrations and platform transformations to rebuilding production infrastructure, including one platform where availability improved from approximately 90% to 99.95%.
AI Systems
I now apply the same engineering principles to AI and agent-based systems. My focus is not simply integrating a model into an application — it is designing the system around it:
- Local and hosted model orchestration
- On-device AI
- Agent workflows
- Evaluation and quality gates
- Automated verification
- Persistent agent state
- Cost and usage controls
- Human approval where it matters
I build systems where model output is treated as something that must be measured and verified rather than automatically trusted. That distinction becomes increasingly important as AI systems move from answering questions to performing work.
How I Work
- Understand the existing system.
- Find the dependencies.
- Design the target state.
- Design the path between them.
- Build the controls that show whether it is actually working.
- And leave behind something that other people can operate.