Flagship course
Enterprise AI Architecture
Design production AI as a platform, not a prompt.
Outcomes
- Draw an enterprise AI architecture that separates models, context, retrieval, agents, and runtimes.
- Choose between workflows and agents with explicit trade-offs.
- Place identity, authorization, and evaluation in the control plane rather than the prompt.
- Produce an ADR-ready reference architecture for a real product line.
Who It Is For
- Enterprise and solution architects
- Staff and principal engineers moving into AI architecture
- Platform leads standing up a shared AI capability
Prerequisites
- Comfort with distributed systems and APIs
- Working familiarity with LLMs in production or prototype form
- How LLMs Work series (free on Learn) recommended
Curriculum
LLM systems
- Model mental models
- Inference economics
- Model selection and routing
Context and retrieval
- Context budgets
- RAG architectures
- Access-aware retrieval
Agents and runtimes
- When not to use an agent
- Harness design
- Durable execution
Trust
- Agent identity
- Authorization at the tool boundary
- Threat models
Platforms and governance
- AI gateways
- Operating model
- Build vs buy
Architecture Labs
Labs focus on architecture decisions, not toy coding exercises.
- Map a copilot onto the architecture layers
- Write an ADR: agent vs workflow
- Design a harness boundary for a privileged tool
Included Resources
Instructor
Hammad Abbasi — AI & Security Architect. 16+ years across identity, distributed systems, and production agent architecture. Speaker at EIC KuppingerCole.