Agentic AI & Automation — Skills Plan
Planned skill map for Agentic AI & Automation. The curriculum will be developed and released in stages.
Design multi-step LLM workflows that call tools, preserve state, recover from failures, and require human approval for consequential actions.
Potential roles: Agentic AI Engineer · AI Automation Engineer
Before you start
- Python
- HTTP APIs and authentication
- LLM APIs and structured outputs
- RAG and vector databases
What you’ll build
- Read-only research agent
- Human-approved operations workflow
- Observable multi-step support assistant
Prerequisites
PLANNED- Python
Build the Python knowledge needed to follow the Agentic AI & Automation foundations.
- HTTP APIs and authentication
Build the HTTP APIs and authentication knowledge needed to follow the Agentic AI & Automation foundations.
- LLM APIs and structured outputs
Build the LLM APIs and structured outputs knowledge needed to follow the Agentic AI & Automation foundations.
- RAG and vector databases
Build the RAG and vector databases knowledge needed to follow the Agentic AI & Automation foundations.
Foundations
PLANNED- Agent loop and workflow patterns
Explain how Agent loop and workflow patterns works and where it fits in the Agentic AI & Automation learning path.
- State, context, and memory
Explain how State, context, and memory works and where it fits in the Agentic AI & Automation learning path.
- Tool/function calling
Explain how Tool/function calling works and where it fits in the Agentic AI & Automation learning path.
- Structured decisions and routing
Explain how Structured decisions and routing works and where it fits in the Agentic AI & Automation learning path.
Core Skills
PLANNED- Planning and task decomposition
Use Planning and task decomposition to implement a working capability and verify its behavior in Agentic AI & Automation.
- Tool schemas and authorization
Use Tool schemas and authorization to implement a working capability and verify its behavior in Agentic AI & Automation.
- Human-in-the-loop approvals
Use Human-in-the-loop approvals to implement a working capability and verify its behavior in Agentic AI & Automation.
- LangGraph or comparable orchestration
Use LangGraph or comparable orchestration to implement a working capability and verify its behavior in Agentic AI & Automation.
- Model Context Protocol (MCP) fundamentals
Use Model Context Protocol (MCP) fundamentals to implement a working capability and verify its behavior in Agentic AI & Automation.
Advanced / Production
PLANNED- Retries, idempotency, and recovery
Evaluate Retries, idempotency, and recovery trade-offs against the performance, reliability, and safety needs of Agentic AI & Automation systems.
- Multi-agent coordination and trade-offs
Evaluate Multi-agent coordination and trade-offs trade-offs against the performance, reliability, and safety needs of Agentic AI & Automation systems.
- Evaluation and deterministic testing
Evaluate Evaluation and deterministic testing trade-offs against the performance, reliability, and safety needs of Agentic AI & Automation systems.
- Tracing, observability, and replay
Evaluate Tracing, observability, and replay trade-offs against the performance, reliability, and safety needs of Agentic AI & Automation systems.
- Guardrails, security, and prompt-injection defense
Evaluate Guardrails, security, and prompt-injection defense trade-offs against the performance, reliability, and safety needs of Agentic AI & Automation systems.
- Budgets, deployment, and incident handling
Evaluate Budgets, deployment, and incident handling trade-offs against the performance, reliability, and safety needs of Agentic AI & Automation systems.
Real-World Projects
PLANNED- Read-only research agent
Deliver a working outcome with Read-only research agent, then test it against the project requirements.
- Human-approved operations workflow
Deliver a working outcome with Human-approved operations workflow, then test it against the project requirements.
- Observable multi-step support assistant
Deliver a working outcome with Observable multi-step support assistant, then test it against the project requirements.
Career / Interview Preparation
PLANNED- Agent architecture review
Explain your approach to Agent architecture review, defend key trade-offs, and discuss what you learned.
- Failure-mode analysis
Explain your approach to Failure-mode analysis, defend key trade-offs, and discuss what you learned.
- Tool security and approval design
Explain your approach to Tool security and approval design, defend key trade-offs, and discuss what you learned.
- Project architecture walkthrough
Explain your approach to Project architecture walkthrough, defend key trade-offs, and discuss what you learned.
Discussion
0Have a better approach? Found something outdated? Share it — your knowledge helps everyone learning here.