Skip to content
Expert Services
Contact Experts

GitHub Copilot prompt engineering

Overview

In this session, GitHub’s experts will provide the participants with the knowledge and techniques necessary to design and refine prompts to elicit the most accurate, relevant and valuable responses from large language model-based systems such as GitHub Copilot.

This course delves into the fundamentals of prompt engineering starting at its origin, exploring the principles of prompt design, the impact of prompt structure, and the importance of context and clarity. Participants will learn how to create precise and targeted prompts, troubleshoot common issues, and apply best practices to ensure optimal outcomes.

By understanding the nuances of AI language models and their response patterns, participants will be equipped to reach the full potential of these technologies, driving improvements in productivity, creativity, and problem-solving within their organization.

Topics

  • Prompt engineering introduction
  • GitHub Copilot model foundations
  • GitHub Copilot backend on security and filters
  • Context in GitHub Copilot
  • Improving prompts with best practices
  • Prompt engineering techniques
  • GitHub Copilot in practice

Customer benefits

The offering will help customers:

  • Design effective prompts that significantly improve the quality and relevance of Copilot-generated responses
  • Utilize models and GitHub Copilot efficiently, understanding the nuances of context windows, prompt construction, and response optimization.
  • Apply proven heuristics and best practices to create precise and targeted prompts.
  • Navigate and implement security and privacy measures in prompt engineering, such as PII filtering and public code matching.

Learning objectives

By the end of this training, learners will be able to:

  • Master the core fundamentals of prompt engineering, including how to design and refine prompts for optimal AI interaction and response accuracy
  • Develop insights into AI models, the context window/size, and how prompts are constructed in GitHub Copilot, including context, rank ordering and differences between Copilot Chat and in-line GitHub Copilot
  • Recognize the process of response generation in AI models
  • Discover heuristics and best practices to enhance the effectiveness of prompts
  • Leverage different concepts for prompt enhancing including but not limited to shot prompting, chain-of-thought prompting, and generated knowledge prompting
  • Recognize GitHub Copilot proxy’s role in PII filtering, public code matching, and security checks to ensure safe and compliant prompt usage

Audience

Required:

  • Developers
  • DevOps Engineers

Delivery details

  • Level: Intermediate [200]
  • Offering type: Training
  • Format: Remote
  • Class size: 15 participant maximum (with some flexibility)

Customer prerequisites

Before this training, the customer needs to have in place:

  • GitHub Enterprise Cloud account
  • GitHub Copilot enabled
  • Compatible IDE with GitHub Copilot
  • GitHub Copilot extension installed on the IDE
  • GitHub Copilot fundamentals training or equivalent experience

How can we help?

Let's build a customized solution that meets all of your needs.

This field is required.
Please enter a valid work email address.
This field is required.
This field is required.
For support questions, head to
github.com/contact
Back to catalog listing