Kurskode: STA_CAI

varighet: 1

Sted: Live Virtual

Kategori: Applications Development

Kurs oversikt

Coding with AI Course Overview

There’s no doubt that AI can make developers more efficient and productive.It’s also a great learning aid, enabling individuals to acquire knowledge intheir own way and at their own pace. But it has its drawbacks too. It’s costly.LLM responses may be of poor quality, wrong, or even malicious. Developers mustknow enough to critique the code produced, and they must be disciplined enough toavoid cognitive atrophy.

This course deals with key AI concepts including language models, prompting,context, and agents.
It provides comprehensive coverage of GitHub Copilot in VS Code specifically,though the concepts are applicable to other AI coding assistants too. Delegateswill learn how to exploit suggestions, agents, context, tools, slash commands,and smart actions. Delegates will further learn how to customise Copilot and touse it safely and securely.

Exercises and examples are used throughout the course to give practicalhands-on experience with the techniques covered.

Course Contents

Course Introduction

  • Administration and course materials
  • Course structure and agenda
  • Delegate and trainer introductions

Session 1: LANGUAGE MODELS AND AGENTS

  • Models
  • Prompts
  • Prompt engineering
  • LLM characteristics
  • Thinking and reasoning
  • Which model?
  • Context
  • Context window
  • Tools
  • What is an agent?
  • The agent loop
  • Agent types
  • Subagents
  • Memory

Session 2: GITHUB COPILOT IN VISUAL STUDIO CODE

  • About GitHub Copilot
  • Setup
  • Suggestions
  • Agents (chat)
  • Prompting
  • Context
  • Tools
  • Reviewing
  • Editing previous chat requests
  • Checkpoints
  • Chat participants
  • Inline chat
  • Generate code context menu
  • Slash commands
  • Smart actions
  • Switching accounts
  • Signing out
  • Disabling AI features

Session 3: COPILOT CUSTOMISATION

  • Reusable prompts
  • Custom instructions
  • Custom agents
  • Model Context Protocol (MCP)

Session 4: COPILOT SECURITY

  • Execution and access
  • Supply chain dependencies
  • Approval automation
  • Information exposure
  • Prompt injection
  • Trust boundaries
  • Controlled scope
  • Transparency
  • Secrets management
  • Recommended security baseline
  • Enterprise policies

Course Objectives

The course aims to provide the delegate with the skills and knowledge necessaryto use AI tools, specifically GitHub Copilot in VS Code, to become moreefficient and productive. It also seeks to provide the delegate with anunderstanding of the way in which LLMs and agents work, and the risks andsafeguards involved.

Requirements

Delegates attending this course will, ideally, be able to write codeindependently. This does not apply to those attending with a view tounderstanding the concepts, risks, and safeguards only.

Who will the Course Benefit?

The Coding with AI course is primarily aimed at developers, analysts, andengineers (those who write code), but it may also be of interest tonon-technical people responsible for teams that produce code.

Kontakt oss: Kurs@sgpartner.no