Artificial intelligence

1. NLU (Natural Language Understanding)

2. DM (Dialog Management)

3. NLG (Natural Language Generation)

4. Tokenization

5. Stop words

6. Lemmatization and stemming

7. Bag of words

8. TF-IDF (term frequency-inverse document frequency)

9. Intent recognition

• Prompt Engineering
• Zero-shot learning
• Few-shot learning
• Fine-tuning
• Retrieval-Augmented Generation (RAG)

• Ollama
• vLLM
• llama.cpp
• LM Studio
• koboldcpp

• Cherry Studio
• Open WebUI
• Msty Studio
• Librechat
• Chatbox
• AnythingLLM
• GPT4All
• Jan.ai
• Text Generation Web UI

• LangChain
• Llamaindex
• Fastmcp
• python-sdk

• Ask Clarifying Questions: Instruct the model to ask follow‑up questions if any part of the prompt is unclear — this prevents misinterpretation and ensures alignment with user intent.
• Provide Examples: Include specific examples to demonstrate the desired output format and style — concrete samples help the model grasp nuances and replicate the expected result.
• Request Reasoning: For complex tasks, ask the model to explain its reasoning step‑by‑step — this increases transparency, allows error tracing, and improves trust in the output.
• Role Definition: Specify the role or persona you want the model to adopt — a defined role (e.g., “expert linguist”) guides tone, depth, and perspective consistently.
• Context Setting: Provide relevant background information to guide the response — context anchors the model’s knowledge and ensures responses are situationally appropriate.
• Define Subtasks: Create specific subtasks for each component — breaking down the prompt into actionable steps helps the model manage complexity and deliver complete answers.
• Identify Main Components: Break the task into logical sections — clear segmentation improves organization and ensures all aspects of the request are addressed systematically.
• Output Structure: Define the structure and format of the desired response — specify headings, bullet points, tables, or paragraph count to match your preferred layout.
• Shadow Prompting: Embed indirect cues and hints within the main prompt to guide the model toward the desired outcome without explicit instructions — subtle framing can shape tone and focus effectively.
• Set Constraints: Define boundaries such as word count, tone, or prohibited topics — constraints prevent irrelevant or overly verbose outputs and keep responses focused.
• Specify Audience: Indicate who the output is for (e.g., beginners, experts) — this adjusts complexity, terminology, and explanatory depth accordingly.
• Use Positive Framing: Frame requests in a constructive way (e.g., “include key benefits” vs. “don’t forget benefits”) — positive phrasing yields more cooperative and complete responses.
• Prioritize Elements: Rank components by importance (e.g., “focus first on X, then Y”) — this guides the model to emphasize critical aspects when trade‑offs arise.
• Include Validation Criteria: State how the output will be evaluated (e.g., accuracy, clarity, creativity) — this aligns the model’s priorities with your quality standards.
• Leverage Iteration: Design prompts to allow follow‑up refinements (e.g., “provide a draft, then we’ll improve it”) — iterative prompting often yields higher‑quality final results.
• Metaprompting: Instruct the model to reflect on its own reasoning or prompt‑design process (e.g., “Explain how you interpreted this prompt” or “Suggest improvements to this request”) — this enhances self‑awareness and helps refine future prompts.
• Prompt Templates: Save effective prompts as reusable templates with placeholders (e.g., [TOPIC], [AUDIENCE]) — this ensures consistency, speeds up workflow, and allows quick adaptation for similar tasks.

• Developing Apps with GPT-4 and ChatGPT (Olivier Caelen, Marie-Alice Blete)
• RAG-Driven Generative AI (Denis Rothman)
• Prompt Engineering for Generative AI: Future-Proof Inputs for Reliable AI Outputs at Scale (James Phoenix & Mike Taylor)