Guides

Each guide collects the articles on one topic, from the basic concept to the production case.

• Prompt engineering fundamentals

  1 article

  See the full guide →

• Example-based prompting

  5 articles

  • What is few-shot prompting?
  • What is one-shot prompting?
  • What is zero-shot prompting?
  • Zero-shot, one-shot or few-shot: which to use

  See the full guide →

• Reasoning and chains of thought

  5 articles

  • What is chain-of-thought (CoT) prompting?
  • Few-shot CoT: reasoning examples in the prompt
  • What is self-consistency in prompting?
  • Zero-shot CoT: 'let's think step by step'

  See the full guide →

• AI agents and agentic prompting

  5 articles

  • Agent, workflow or chatbot?
  • What changes in an agentic prompt
  • The agent loop, step by step
  • What is tool use?

  See the full guide →

• MCP — Model Context Protocol

  5 articles

  • MCP architecture: host, client and server
  • Tools, resources and prompts in MCP
  • MCP transports: stdio and HTTP
  • Why MCP exists: the N×M problem

  See the full guide →

• RAG, retrieval and memory

  4 articles

  • Chunking: strategies and sizes
  • What is an embedding?
  • RAG architecture from scratch

  See the full guide →

• Language models, families and benchmarks

  4 articles

  • What is a foundation model? Base vs instruct
  • How an LLM is trained: pretraining, SFT, alignment
  • What is a transformer? The architecture explained

  See the full guide →

• Prompt and LLM application security

  4 articles

  • Direct or indirect injection?
  • What is an LLM jailbreak?
  • What is prompt injection?

  See the full guide →

• Optimization, evaluation and observability

  4 articles

  • What is DSPy?
  • How DSPy compiles prompts
  • Tune prompts by hand or automate?

  See the full guide →

• Fine-tuning and alignment

  4 articles

  • What is LoRA: low-rank adaptation
  • What is SFT (supervised fine-tuning)?
  • When fine-tuning is worth it

  See the full guide →

• Context and context engineering

  4 articles

  • Context or prompt engineering?
  • What is a model's context window?
  • Lost in the middle: the forgotten middle

  See the full guide →

• Multimodal AI and visual prompting

  1 article

  See the full guide →

• Structured outputs and integration

  4 articles

  • How to force valid JSON from an LLM
  • Grammar-constrained decoding
  • Structured outputs or function calling?

  See the full guide →

• Applications and use cases

  1 article

  See the full guide →

• AI fundamentals

  1 article

  See the full guide →

• Research and how the field evolves

  1 article

  See the full guide →

• How models work under the hood

  4 articles

  • The embedding and unembedding layers
  • From text to tokens: the input path
  • What is the residual stream

  See the full guide →

• Model generations and changelogs

  1 article

  See the full guide →

• Evaluation and capability measurement

  1 article

  See the full guide →

• Agent harness

  4 articles

  • Context management in the agent harness
  • The layers of an agent harness
  • Automatic agent session compaction

  See the full guide →

• The frontier and what's still unsolved

  1 article

  See the full guide →