Get in Touch
 Duration 14 hours

Course Outline

Introduction

  • Distinguishing between workflows and agents
  • High-level architecture overview and common use cases

Claude API Fundamentals

  • The Messages API, available models, and key parameters
  • Crafting system prompts and managing structured output
  • Handling streaming responses and errors
  • Lab: Integrating the API for the first time

Tool Utilisation

  • Defining tools and their schemas
  • The tool-use loop and strategies for error recovery
  • Lab: Creating an agent that interacts with internal company APIs

MCP and Agent Skills

  • Understanding MCP servers and clients
  • Leveraging existing servers and developing a basic custom one
  • Implementing Agent Skills for reusable expertise
  • Lab: Connecting an agent to an external data source

Claude Agent SDK

  • The agent loop, session management, and context handling
  • Managing permissions, hooks, and subagents
  • Lab: Building a research and reporting agent

Cost, Quality, and Safety

  • Implementing prompt caching and token budgeting
  • Establishing guardrails and human approval workflows
  • Conducting evaluations with a small test set
  • Implementing logging and monitoring practices

From Prototype to Production

  • Deployment strategies and secure secret management
  • Ongoing maintenance and version control
  • Developing a roadmap for your specific use case

Requirements

  • Proficiency in Python or TypeScript
  • Practical experience with REST APIs and JSON handling
  • A foundational understanding of large language models

Target Audience

  • Software developers and data engineers integrating AI features into products and internal tools
  • Solution architects assessing agent-based architectures
  • Technical teams aiming to transition from prototype stages to production environments

Number of participants


Price per participant

Testimonials (3)

Provisional Upcoming Courses (Require 5+ participants)

Related Categories