🤖 AI & Machine Learning
Generative AI: Prompt Engineering Basics
This course builds a practical foundation in prompt engineering for large language models, taking you from core concepts...
69
Views
Sections
1. Foundations of Generative AI2 views
Establish how modern LLMs generate text, the role of tokens and probabilities, and the constraints that shape prompt behavior.
15 topics (15 versions)
1.1What Is Generative AI
1
1.2AI vs ML vs Deep Learning
1
1.3Transformer Architecture Primer
1.4Tokens and Tokenization
1.5Probabilities and Next-Token Prediction
1.6Temperature and Top-p Sampling
1.7Context Window and Limits
1.8Prompt–Response Loop
1.9System, Developer, and User Messages
1.10Capabilities and Limitations
1.11Hallucinations and Uncertainty
1.12Determinism vs Stochasticity
1.13Safety Layers and Moderation
1.14Evaluation Mindset from Day One
1.15Useful Mental Models of LLMs
2. LLM Behavior and Capabilities3 views
Understand alignment, sensitivity to phrasing, non-determinism, and other behavioral properties that your prompts must account for.
15 topics (15 versions)
2.1Pretraining and Fine-Tuning
2.2Instruction Following and Alignment
2.3RLHF and Preference Optimization
2.4Sensitivity to Wording and Order
2.5Length Bias and Cutoff Realities
2.6Hidden Biases and Stereotypes
2.7Refusals and Safety Behavior
2.8Non-Determinism and Sampling Variance
2.9Stop Sequences and Output Control
2.10System Message Priority
2.11Tool-Use Affordances
1
2.12Function Calling at a Glance
2.13Style and Tone Emulation
1
2.14Domain Transfer and Generalization
1
2.15When Models Say “I Don’t Know”
3. Core Principles of Prompt Engineering11 views
Adopt guiding principles—clarity, specificity, grounding, and iteration—to consistently steer models toward desired outcomes.
15 topics (15 versions)
3.1Clarity Over Cleverness
2
3.2Specificity and Constraints
2
3.3User Intent and Task Framing
1
3.4Audience and Tone Control
1
3.5Context and Grounding
3.6Example-Driven Guidance
1
3.7Output Structure and Formatting
1
3.8Stepwise Reasoning Prompts
3.9Verification and Fact-Checking
1
3.10Controlling Randomness
1
3.11Guardrails and Boundaries
3.12Assumption Surfacing
3.13Decomposition Before Execution
1
3.14Iteration and Refinement Cycles
3.15Success Criteria Up Front
4. Writing Clear, Actionable Instructions0 views
Craft precise directives with scope, constraints, and acceptance criteria that remove ambiguity and reduce rework.
15 topics (15 versions)
4.1Choose Strong Action Verbs
4.2Define Scope and Boundaries
4.3State Acceptance Criteria
4.4Include Constraints and Limits
4.5Numbered Steps and Checklists
4.6Avoid Ambiguity and Vague Terms
4.7Use Negative Prompts Sparingly
4.8Disclose Time and Context
4.9Domain Vocabulary and Glossaries
4.10Reference Styles and Citations
4.11Multi-Task Prompt Patterns
4.12Question Framing Techniques
4.13Brevity vs Completeness
4.14Hints and Nudge Strategies
4.15Avoid Leading the Model
5. Roles, Personas, and System Prompts0 views
Leverage roles and system instructions to shape expertise, tone, and boundaries across single and multi-agent setups.
15 topics (15 versions)
5.1Selecting Effective Roles
5.2Calibrating Expertise Levels
5.3Voice, Style, and Tone
5.4Constraint-Driven Personas
5.5Multiple Personas in Dialogue
5.6Role-Based Guardrails
5.7Hierarchical Prompting Patterns
5.8Chain-of-Roles Workflows
5.9Maintaining Persona Consistency
5.10Audience Emulation Prompts
5.11System vs Developer vs User
5.12Prefix and Header Templates
5.13Style Guides as Prompts
5.14Refusal and Safety Personas
5.15Persona Handoffs and Transitions
6. Supplying Context and Grounding0 views
Feed the model the right facts at the right time using structured context blocks, delimiters, and source pinning.
15 topics (15 versions)
6.1Curating Background Information
6.2Injecting Data Snippets
6.3Grounding With Sources
6.4Retrieval Summaries in Prompts
6.5Citing and Linking Evidence
6.6Planning Context Budgets
6.7Chunking and Windowing
6.8Pinning Critical Facts
6.9Canonical Source Selection
6.10Structured Context Blocks
6.11Delimiters and Separators
6.12Unknowns and Clarification Triggers
6.13Session Memory Strategies
6.14Preventing Context Leakage
6.15Updating Stale Context
7. Examples: Zero-, One-, and Few-Shot0 views
Use demonstrations to steer behavior, balancing exemplar quality, order effects, and when to skip examples entirely.
15 topics (15 versions)
7.1When to Use Zero-Shot
7.2One-Shot Demonstrations
7.3Few-Shot Prompt Patterns
7.4Selecting Quality Exemplars
7.5Counterexamples for Boundaries
7.6Order and Primacy Effects
7.7Formatting Exemplars Cleanly
7.8Label and Schema Consistency
7.9Difficulty Gradient Design
7.10Input–Output Pair Crafting
7.11Contrastive Example Pairs
7.12Minimal Pair Construction
7.13Avoiding Selection Bias
7.14Decision Boundary Illustration
7.15Knowing When to Skip Examples
8. Structuring Outputs and Formats0 views
Specify output schemas, enforce structure, and design responses for easy parsing, scoring, and downstream use.
15 topics (15 versions)
8.1Bullets, Lists, and Outlines
8.2Headings and Sections
8.3JSON and Schema Enforcement
8.4XML, CSV, and TSV Outputs
8.5Markdown and Code Blocks
8.6Tags, Markers, and Delimiters
8.7Length Targets and Summaries
8.8Scoring Rubrics and Scales
8.9Key-Value Output Patterns
8.10Placeholders and Variables
8.11Error Message Conventions
8.12Structured Reasoning Fields
8.13Tables and Matrices
8.14Multi-Part Output Assembly
8.15Post-Processing Friendly Designs
9. Reasoning and Decomposition Techniques0 views
Elicit better thinking with outline-first strategies, hypothesis testing, and verification-first prompting.
15 topics (15 versions)
9.1Outline-Then-Detail Pattern
9.2Scratchpad and Notes Fields
9.3Rationale-Lite Approaches
9.4Self-Ask and Subquestioning
9.5Hypothesis Generation
9.6Back-Solving Strategies
9.7Plan-Then-Execute Split
9.8Compare-and-Contrast Prompts
9.9Constraint Propagation
9.10Uncertainty and Confidence Cues
9.11Verification Steps First
9.12Sanity Checks and Estimation
9.13Socratic Questioning Prompts
9.14Eliminating Irrelevant Paths
9.15Chain-of-Thought Considerations
10. Iteration, Testing, and Prompt Debugging0 views
Develop a rigorous workflow to test, analyze, and refine prompts using experiments, versioning, and red teaming.
15 topics (15 versions)
10.1Test Case Design
10.2A/B and Multivariate Prompt Tests
10.3Minimal Reproducible Prompts
10.4Error Pattern Analysis
10.5Prompt Ablation Studies
10.6Parameter Sweep Experiments
10.7Red Teaming for Robustness
10.8Guardrail Trigger Testing
10.9Fallback and Recovery Prompts
10.10Versioning and Naming Conventions
10.11Change Logs and Diffing
10.12Regression Test Suites
10.13Canary Questions and Probes
10.14Peer Review and Pair Prompting
10.15Capturing Learnings and Playbooks
11. Evaluation, Metrics, and Quality Control0 views
Measure output quality with human and automated methods, track performance, and close the loop with monitoring.
15 topics (15 versions)
11.1Human Evaluation Rubrics
11.2LLM-as-Judge Techniques
11.3Objective vs Subjective Metrics
11.4Accuracy, Fluency, and Coverage
11.5Safety and Harms Assessment
11.6Cost, Latency, and Quality Tradeoffs
11.7Acceptance Thresholds
11.8Inter-Rater Reliability
11.9Sampling and Test Sets
11.10Calibration and Score Normalization
11.11Prompt Scorecards
11.12Dashboards and Monitoring
11.13Continuous Evaluation Loops
11.14Drift and Degradation Detection
11.15Closing the Feedback Loop
12. Safety, Ethics, and Risk Mitigation0 views
Build safe prompts that reduce harm, protect privacy, handle sensitive content, and maintain accountability.
15 topics (15 versions)
12.1Harmful Content Avoidance
12.2Bias and Fairness Controls
12.3Privacy and PII Handling
12.4Copyright and Licensing
12.5Hallucination Containment
12.6Verification Before Action
12.7Domain-Specific Risk Patterns
12.8Prompt Injection Awareness
12.9Jailbreak Resistance Strategies
12.10Secure Delimiters and Sandboxing
12.11Sensitive Topic Handling
12.12Consent and User Safeguards
12.13Age-Appropriate Design
12.14Transparency and Disclosure
12.15Accountability and Audit Trails
13. Tools, Functions, and Agentic Workflows15 views
Integrate function calling and tools, design planner–executor patterns, and manage errors, scopes, and observability.
15 topics (15 versions)
13.1Function Calling Patterns
2
13.2Parameter Schema Design
1
13.3Tool Selection Prompts
1
13.4Planner–Executor Architectures
13.5Grounding via External Tools
1
13.6Error Handling and Retries
1
13.7Timeouts and Circuit Breakers
1
13.8Result Summarization Prompts
1
13.9Chaining Tool Calls
1
13.10Calculators, Coders, and Browsers
1
13.11Tool Availability Checks
1
13.12Permissions and Scopes
1
13.13Semantic Caching Strategies
1
13.14Observability and Logs
1
13.15Fallback to Tool-Free Modes
1
14. Retrieval-Augmented Generation (RAG)0 views
Combine prompts with retrieval to ground answers in external knowledge, improving accuracy and traceability.
15 topics (15 versions)
14.1RAG Concepts and Benefits
14.2Embeddings and Vectorization
14.3Indexing and Chunking Tactics
14.4Query Construction Prompts
14.5Re-Ranking and Fusion
14.6Citation and Attribution Formats
14.7Freshness and Recency Strategies
14.8Reducing Hallucinations with RAG
14.9Hybrid Sparse–Dense Search
14.10Context Compression Techniques
14.11Budget-Aware Retrieval
14.12RAG Evaluation Methods
14.13Answer–Source Separation
14.14Dynamic Routing and Switching
14.15Vector Store Hygiene
15. Multimodal and Advanced Prompt Patterns2 views
Extend prompting across text, images, audio, and code while adopting emerging patterns and deployment guardrails.
15 topics (15 versions)
15.1Image–Text Prompting
15.2Audio and Speech Prompts
15.3Code Generation Prompts
1
15.4Agent and Orchestrator Patterns
15.5Collaborative Prompting Workflows
1
15.6Meta-Prompts and Self-Reflection
15.7Ensemble and Voting Prompts
15.8Time- and Date-Aware Prompts
15.9Multilingual and Translation Prompts
15.10Cultural and Style Adaptation
15.11Long-Context Prompting
15.12Session Memory Management
15.13Template Libraries and Snippets
15.14Deployment Guardrails
15.15Emerging Trends and Research