Anthropic Claude

Anthropic Claude is a family of frontier large language models — Opus, Sonnet, and Haiku in the current Claude 4.X generation — and a growing surface of features around it: a 200K to 1M-token context window depending on tier, prompt caching, tool use and parallel tools, computer use, vision, PDF / document handling, citations, the Claude Agent SDK, the Claude Code CLI, the Model Context Protocol (MCP), the Batch API, and managed agents. Over the last two years Claude has stopped being a research demo and become an operating choice for product teams: every meaningful AI feature in a modern application now sits behind a model-selection decision, a prompt-engineering decision, a tool-use design decision, an evaluation decision, and a production-cost decision. Practitioners who understand the family and the surface ship better products faster, debug more cleanly, and waste less money on the gap between “works in a notebook” and “runs in production”.

This track is written for the practitioners building with Claude day to day: ML engineers, software engineers integrating LLM features, product managers shipping AI capabilities, prompt engineers, agent builders, T&S teams using Claude for moderation, RAI leads writing evaluations, security engineers reviewing LLM deployments, and platform leads choosing between model providers. Every topic explains the underlying capability or feature (drawing on Anthropic’s public documentation, system cards, the Anthropic SDK source, the canonical research the model is built on, and hard-won production experience), the practical pattern that operationalises it, and the failure modes that quietly trip teams up. The aim is that a reader can build with Claude credibly, defend the choice to a sceptical engineering review, and run the resulting system in production. This course content is independent educational material; Lilly Tech Systems is not affiliated with Anthropic, and Anthropic, Claude, and related marks are the property of their respective owners.