17 free courses, no signup wall
Architect-led enterprise cloud, security & AI
Fixed-price engagements, scoped on a discovery call
Skip to content
Agent engineer

AI Agent Engineering

Agents are principals, not features. Each one needs an identity, a scope, and an audit trail.

Who this is for

Engineers and architects putting autonomous or semi-autonomous agents into a real environment.

What you should be able to do

Run an agent under a named autonomy tier with scoped credentials and an attributable log.

Career ladder

The titles this track maps onto. Levels differ between companies. The useful part is the direction, and what each step adds to the one before it.

  1. Automation Engineer
  2. Agent Engineer
  3. Senior Agent Engineer
  4. Staff Agent Engineer
  5. Principal Agent Architect

Tech stack

What the work is actually done with. Grouped by the job each tool does, so the list reads as a system rather than a pile of names.

Orchestration
LangGraphMCPBedrock AgentsVertex AI Agents
Memory
Vector storeEpisodic storeSummarisation
Tools
Tool callingMCP serversWebhooks
Control
Autonomy tiersHuman approvalScoped credentialsAudit log

The delivery flow

The order the work actually happens in. Each step is where a decision gets made and written down, not a chapter heading.

  1. Task definition
  2. Agent identity
  3. Scoped credentials
  4. Tool registry
  5. Planner / orchestrator
  6. Autonomy tier gate
  7. Human approval
  8. Audit trail

Reference repository structure

A starting layout for this track. The directories are the ones that get added late and hurt. Decisions, evals, policy, lineage. Promoted to the top level where they are visible.

agent-platform/
├── agents/
│   ├── planner/
│   ├── researcher/
│   ├── reviewer/
│   └── orchestrator/
├── tools/                 # one module per tool, schema + handler
├── mcp-servers/           # capability boundaries, one per surface
├── memory/
│   ├── working/
│   └── episodic/
├── policy/
│   ├── autonomy-tiers.yaml
│   └── approval-rules.yaml
├── identity/              # per-agent credentials, scoped and rotated
├── audit/                 # append-only, attributable to one agent
└── evals/

Reading path

12 published guides on this track. Every one is a live page on this site. Nothing here is a placeholder.

  1. What Is Agentic AI? A Practical Enterprise Definition
  2. What Is the Model Context Protocol (MCP)?
  3. Autonomy Tiers for AI Agents: A Classification and Enforcement Guide
  4. A Phased Governance Rollout for AI Agents: Crawl, Walk, Run
  5. AI Agent Security: Threats and Controls
  6. AI Agent Services, Build Production AI [2026]
  7. AI Agents in Enterprise: From POC to Production
  8. AI Engineering Course: LLMs, RAG, Agents [2026]
  9. Audit Trails for AI Agents: Making Every Action Attributable
  10. Enterprise AI Security: Protecting LLMs, Agents & Data Pipelines [2026]
  11. How to Get AI Agents Through Enterprise Security Review
  12. Kubernetes for Beginners 2026: A Plain-English Guide to Container Orchestration

Reference implementations

Citadel’s open-source repositories for this track. Terraform modules, MCP servers, and reference architectures you can read, fork, and run. Apache/MIT licensed; check each repository for its terms.

Primary sources

The standards, framework documents, and vendor references this track is built against. Go here when you need the authoritative wording rather than a summary. In a security review or an audit, the source is what counts.

Other tracks

Talk through this with an architect