How to Reduce Hallucinations in LLM Applications
Proven ways to reduce LLM hallucinations: grounding with RAG, citations, structured output, abstention, verification checks, and evals that catch errors early.
WRITING / NOTES FROM THE FORGE
AI
12 articles tagged AI.
Fine-Tuning vs RAG vs Prompt Engineering: Which Should You Use?
Fine-tuning vs RAG vs prompt engineering compared: what each changes, costs, and fixes, plus a simple decision framework for choosing the right LLM approach.
Prompt Injection: What It Is and How to Defend Your AI App
Prompt injection explained: direct vs indirect attacks, why no complete fix exists, and the layered defenses that limit damage in LLM apps and AI agents.
LLM Evaluation: How to Test AI Features Before Production
A practical guide to LLM evaluation: build a golden dataset, pick metrics, use LLM-as-judge carefully, and gate releases so AI features do not regress.
Model Context Protocol (MCP) Explained With Examples
What is MCP? The Model Context Protocol explained with a real server example, message flow, tools vs resources vs prompts, transports, and security tips.
How to Build an AI Agent: Architecture, Tools, and Guardrails
Learn how to build an AI agent step by step: the core loop, tool design, memory, and guardrails that keep an LLM agent useful, safe, and cheap in production.
Embeddings Explained: How AI Understands Meaning
Embeddings explained for developers: what embeddings are, how they capture meaning, how cosine similarity works, and how to use them for search and RAG.
How to Choose a Vector Database for Your AI Application
How to choose a vector database: a practical comparison of dedicated vector DBs, pgvector, search engines, and libraries by scale, filtering, and cost.
RAG Explained: How to Build Retrieval-Augmented Generation That Works
RAG explained for developers: how retrieval-augmented generation works, how to chunk, embed, retrieve, and rerank, and how to evaluate a RAG pipeline.
Prompt Engineering Best Practices for Developers
Prompt engineering best practices for developers: clear instructions, examples, structured output, delimiters, evals, and patterns that make LLMs reliable.
How to Reduce LLM API Costs: 12 Proven Techniques
Learn how to reduce LLM API costs with 12 proven techniques, from prompt caching and model routing to batching, semantic caching, and token monitoring.
What Is a Large Language Model? A Plain-English Guide for Developers
What is a large language model? A plain-English guide for developers covering tokens, transformers, training, context windows, and how to use LLMs well.