Skip to content

Prompt Engineering

Title: Prompt Engineering: From Fundamentals to Agentic AI

Target Audience: General audience with at least a high school education, including business professionals, educators, students, and anyone seeking to effectively communicate with large language models (LLMs). No prior programming or AI experience is required.

Prerequisites: None. Participants should have basic computer literacy and access to a web browser. Familiarity with any AI chatbot (such as ChatGPT, Claude, or Gemini) is helpful but not required.

Course Overview

Prompt engineering is the practice of crafting effective instructions for large language models to produce accurate, relevant, and useful responses. As AI becomes deeply integrated into workplaces, education, and daily life, the ability to communicate precisely with these systems has become an essential modern literacy. This course provides a comprehensive, hands-on introduction to prompt engineering that empowers participants to harness AI tools effectively and responsibly.

This course covers the full spectrum of prompt engineering, from foundational principles like clarity and specificity to advanced techniques such as chain-of-thought reasoning, few-shot learning, retrieval-augmented generation (RAG), and GraphRAG — which uses knowledge graphs to answer questions that depend on relationships between entities. Participants will also explore the latest developments in agentic AI, where language models can use tools, execute multi-step workflows, and collaborate with specialized skills to accomplish complex tasks autonomously.

Whether you are a business professional looking to streamline operations, an educator designing AI-enhanced curricula, or a curious learner exploring the capabilities of modern AI, this course will give you practical skills you can apply immediately. All concepts are taught through hands-on labs with real AI systems, ensuring participants leave with confidence and competence.

Main Topics Covered

  • Fundamentals of prompt design: clarity, specificity, and structure
  • Types of prompts: open-ended, closed-ended, instructional, and conversational
  • Prompt patterns: zero-shot, one-shot, and few-shot learning
  • Chain-of-thought and step-by-step reasoning prompts
  • Role and persona assignment in prompts
  • Output format control: markdown, JSON, CSV, tables, and structured data
  • Iterative prompt refinement and A/B testing
  • System prompts, context windows, and token management
  • Token budgets, rate limits, and usage monitoring: understanding five-hour usage windows, estimating token costs, and working within platform limits
  • Retrieval-augmented generation (RAG), GraphRAG, and grounding prompts in data using knowledge graphs
  • Multimodal prompting: working with images, documents, and code
  • Agentic AI: tool use, skills, and multi-step autonomous workflows
  • AI agents and orchestration: planning, execution, and feedback loops
  • Prompt security: avoiding injection attacks and adversarial manipulation
  • Ethical considerations: bias, hallucination detection, and responsible AI use
  • Business applications: customer service, content creation, data analysis, and decision support
  • Educational applications: personalized tutoring, assessment design, and curriculum development

Topics Not Covered

  • Deep learning theory, neural network architecture, or model training
  • Fine-tuning or pre-training language models
  • Programming language-specific AI development (e.g., building ML pipelines)
  • Hardware infrastructure for AI deployment
  • Mathematical foundations of transformer models
  • Specific vendor certifications or proprietary platform administration

Learning Outcomes

After completing this course, students will be able to:

Remember

Retrieving, recognizing, and recalling relevant knowledge from long-term memory.

  • List the key principles of effective prompt design (clarity, specificity, context, format)
  • Identify the major prompt patterns: zero-shot, one-shot, few-shot, and chain-of-thought
  • Recall the components of a well-structured prompt (role, task, context, constraints, output format)
  • Name common prompt engineering pitfalls such as ambiguity, leading questions, and overcomplication
  • Recognize the differences between system prompts, user prompts, and assistant responses
  • Recall common platform usage limits such as rate limits, token budgets, and five-hour usage windows

Understand

Constructing meaning from instructional messages, including oral, written, and graphic communication.

  • Explain why prompt structure and specificity affect the quality of AI responses
  • Describe how context windows and token limits influence prompt design decisions
  • Explain how token budgets, rate limits, and five-hour usage windows affect prompt strategy
  • Summarize the differences between zero-shot, few-shot, and chain-of-thought prompting strategies
  • Interpret how role assignment and persona prompts shape model behavior and output tone
  • Explain the concept of retrieval-augmented generation (RAG) and GraphRAG, and why grounding matters
  • Describe how GraphRAG uses knowledge graphs to answer relationship-based questions that traditional RAG cannot
  • Describe how AI agents use tools, skills, and planning to accomplish multi-step tasks

Apply

Carrying out or using a procedure in a given situation.

  • Write clear, well-structured prompts that produce accurate and relevant AI responses
  • Use few-shot examples to guide model output for specific tasks and formats
  • Apply chain-of-thought prompting to solve multi-step reasoning problems
  • Construct prompts that specify output format (markdown, JSON, tables, lists)
  • Use iterative refinement to improve prompt quality based on model responses
  • Craft system prompts that establish roles, constraints, and behavioral guidelines
  • Design prompts for multimodal inputs including images, documents, and code
  • Estimate token usage for prompts and responses to stay within budget constraints

Analyze

Breaking material into constituent parts and determining how the parts relate to one another and to an overall structure or purpose.

  • Compare the effectiveness of different prompt strategies for a given task
  • Analyze why a prompt produces unexpected or low-quality results and identify root causes
  • Distinguish between hallucinated content and factually grounded responses
  • Examine how prompt structure choices affect bias, tone, and accuracy in outputs
  • Break down complex tasks into sequences of prompts suitable for agentic workflows
  • Evaluate the tradeoffs between prompt complexity and response quality
  • Analyze token usage patterns to identify opportunities for cost reduction

Evaluate

Making judgments based on criteria and standards through checking and critiquing.

  • Assess the quality of AI-generated responses against defined accuracy and relevance criteria
  • Critique prompts for potential security vulnerabilities such as prompt injection
  • Judge whether an AI response contains hallucinated or unsupported claims
  • Evaluate the ethical implications of prompt design choices, including bias and fairness
  • Rate the effectiveness of different prompt engineering approaches for business use cases
  • Appraise when to use simple prompting versus agentic multi-step workflows

Create

Putting elements together to form a coherent or functional whole; reorganizing elements into a new pattern or structure.

  • Design a complete prompt engineering workflow for a real-world business or educational scenario
  • Create a library of reusable prompt templates tailored to specific domains and tasks
  • Build multi-step prompt chains that decompose complex problems into manageable sub-tasks
  • Develop evaluation rubrics for assessing prompt quality and AI output effectiveness
  • Construct an agentic workflow that combines tool use, planning, and iterative refinement
  • Design a prompt security strategy that mitigates injection attacks and adversarial inputs
  • Produce a capstone project demonstrating end-to-end prompt engineering for a chosen application