AI and Large-Language-Model Foundations

Distinguish deterministic software, predictive machine learning, generative AI, and agentic systems by their behavior and risks.

Select among prompting, retrieval-augmented generation, fine-tuning, and model training for a use case based on knowledge freshness, proprietary data, cost, latency, and control needs.

Identify hallucination, context, bias, privacy, security, latency, and cost failure modes in an LLM feature.

Define offline and online evaluation criteria for an LLM-powered user journey.

Estimate the principal latency and cost drivers of an LLM application.