Anatomy of a Prompt — System, User, and Assistant ExplainedUnderstand how system, user and assistant messages shape LLM behavior.8 min read·blog.pragmaticbyharsh.comFeb 15
Choosing Embedding Models and Dimensions: Why 1536 Isn't Always Better Than 384Learn how to choose embedding models and dimensions for production RAG systems. Compare OpenAI, Voyage AI, and Google's free options for embedding11 min read·blog.pragmaticbyharsh.comFeb 10
What Are Embeddings and How Vector Similarity Actually WorksLearn what embeddings are, how vector similarity works, and why understanding magnitude vs direction matters for semantic search and RAG systems.14 min read·blog.pragmaticbyharsh.comFeb 8
How Tokenization Works: BPE and the Algorithm Behind Your LLMLearn how Byte Pair Encoding (BPE) actually works — the algorithm that powers GPT, Claude, and LLaMA tokenizers. Step-by-step with examples.9 min read·blog.pragmaticbyharsh.comFeb 3
What Are Tokens and Why Your LLM Bill Depends on ThemLearn what tokens really are, why they're not words, and how understanding tokenization saves you money on LLM API costs.9 min read·blog.pragmaticbyharsh.comFeb 1