Course

Agent Systems Architecture

Orchestration, memory, tools, and runtimes that survive production.

AdvancedEngineers and architects building production agent systems.Coming soon — self-paced, cohort, or hybrid

Outcomes

Specify an agent loop with stop conditions, approvals, and failure handling.
Separate context, memory, and workflow state.
Design a runtime that owns tools, retries, and sandboxing.
Coordinate multiple agents without unconstrained shared memory.

Who It Is For

AI engineers
Platform engineers
Architects owning agent products

Prerequisites

Building Enterprise AI Agents (free on Learn)
Context Engineering guide

Curriculum

Agent design

Agent vs workflow
Autonomy levels
Human gates

Orchestration

Planner/executor
Supervisor/worker
Handoffs

Runtime

Harness
Checkpoints
Idempotency

Multi-agent

Specialists
Shared state
Coordination protocols

Architecture Labs

Labs focus on architecture decisions, not toy coding exercises.

Decompose a monolith agent
Design durable execution for a long-running task

Included Resources

DiagramsChecklistsLabs

Instructor

Hammad Abbasi — AI & Security Architect. 16+ years across identity, distributed systems, and production agent architecture. Speaker at EIC KuppingerCole.