LLM Concepts
How large language models work and how to use them effectively.
- Fine-Tuning
Engineer LLM fine-tuning with clear prompting and RAG boundaries, governed datasets, LoRA and QLoRA, evaluation gates, and production operations.
- Prompt Engineering
Design, version, and evaluate LLM prompts for production — system vs user roles, few-shot, chain-of-thought, and the limits that push you toward RAG, tools, and fine-tuning.
- Model Context Protocol (MCP)
Engineer MCP host–client–server integrations: tools/resources/prompts, transports, auth, timeouts, and how MCP standardizes tool I/O without replacing orchestration or guardrails.
- Tool Calling
Understand how LLMs invoke external tools and APIs to extend their capabilities beyond text generation.
- Function Calling
Engineer typed LLM tool intents with schema validation, authorization, timeouts, idempotency, observability, and safe production execution.
- Context Windows
Engineer reliable LLM context budgets across instructions, evidence, history, tools, and output—with retrieval, truncation, summarization, and production observability.
- Structured Outputs
Why free-text LLM output fails for machines, constrained decoding vs prompt-only JSON, OpenAI JSON Schema, Pydantic validation, and production patterns with Instructor-style retries.
- Hallucination Detection
Detecting when LLMs fabricate facts — grounding checks, NLI verification, citation validation, and production detection layers.
- Hallucinations
Why LLMs hallucinate — next-token prediction is not truth-seeking — and how to mitigate with RAG, tools, structured outputs, abstention, detection, and verification.