Module 1: Tool & Function Calling
Define typed tools, validate inputs and outputs, handle permissions, and connect model decisions to controlled application actions.
Course Specialization
EnrollAI Agent Development
Program Overview
Build tool-using AI applications with explicit workflows, state, memory, system integrations, guardrails, evaluation, observability, and human review.
Skills & Tools
- Tool & Function Calling
- Workflow Planning
- State & Memory
- System Integration
- Agent Patterns
- Guardrails
- Evaluation & Observability
Course Curriculum
Module 2: Workflow Planning
Model deterministic and model-assisted workflows, break goals into bounded steps, and design explicit stop and recovery conditions.
Module 3: State & Memory
Manage session state, durable task data, retrieval-backed memory, retention boundaries, and user-controlled context.
Module 4: API & Business-System Integration
Integrate approved APIs and business systems with authentication, authorization, idempotency, validation, and audit-friendly records.
Module 5: Single-Agent & Multi-Agent Patterns
Compare orchestration patterns and select the simplest architecture that meets a use case without overstating autonomous behavior.
Module 6: Guardrails & Human Review
Add policy checks, scoped permissions, confirmations, escalation paths, and meaningful review before consequential actions.
Module 7: Evaluation & Observability
Create scenario tests, trace tool use, measure reliability and latency, review failures, and monitor changes over time.
Module 8: Agent Deployment Capstone
Deploy a bounded agent workflow with secure tools, durable state, evaluation evidence, operational monitoring, and a human fallback.
Capstone Projects
Operations Support Agent
Build a permission-aware agent that gathers context, proposes actions, and requires review before updating a connected system.
Agent Deployment Capstone
Deliver and evaluate a tool-using workflow with durable state, observability, failure recovery, and documented operating limits.