Prompt Engineering Patterns
Reusable prompting moves for task framing, constraints, examples, and output shape.
Core lesson
Building With AI
How AI engineers actually build: prompts, agents, retrieval, and evals.
All articles · 17
- 01Tool Use And Function Callingfoundation5 min read
- 02Agent Loopsintermediate4 min read
- 03Model Context Protocolintermediate5 min read
- 04Context Engineeringintermediate5 min read
- 05Vector Embeddingsintermediate3 min read
- 06Semantic Spaceintermediate4 min read
- 07Vector Searchintermediate3 min read
- 08Vector Databasesintermediate3 min read
- 09Hybrid Searchintermediate3 min read
- 10Re-rankingintermediate3 min read
- 11Query Expansionintermediate3 min read
- 12Parent-Child Retrievalintermediate3 min read
- 13Multi-Vector Retrievalintermediate3 min read
- 14Context Compressionintermediate3 min read
- 15Knowledge Graph RAGintermediate3 min read
- 16Agentic RAGintermediate4 min read
- 17LLM Evaluationintermediate4 min read
Planned
Coming soonThese will slot into the reading order above as they're published.
Agent Memory
How agent systems preserve useful state without confusing storage, context, and long-term memory.
Planning And Reasoning Patterns
How AI workflows break tasks into steps, inspect progress, and decide when to stop.
Multi-Agent Systems
When multiple AI workers help, when they add noise, and what coordination costs appear.
Human-In-The-Loop
How approval, review, escalation, and correction keep AI systems under human control.
Agentic AI In Plain English
A beginner-facing explanation of what people mean by agentic AI and what the phrase does not guarantee.
Harness Engineering In Plain English
The software wrapper around a model: prompts, tools, state, policy, traces, and evals.
Document Ingestion And Chunking
How documents become searchable chunks before retrieval and RAG can work well.
LLM-As-Judge And Eval Datasets
How model-based judging and curated cases help evaluate AI behavior without replacing human calibration.