Claude On Google Cloud

Clod Course Overview: Key Concepts

Claude Model Families: Key Concepts

Accessing and Using Claude: Request Lifecycle

I. Setup and Initialization

API Limitation:** The Anthropic API and Cloud do not store any messages (user inputs or model responses).

Chatbot Implementation Concepts

I. Purpose of Customization

Purpose of System Prompts:** System prompts are used to guide the behavior and output style of an AI model.

Text Generation Process

Key Concepts: Implementing Streaming for Chat Interfaces

Influencing LLM Output: Advanced Techniques

Key Concepts: Controlling LLM Output with Stop Sequences and Pre-filling

Study Notes: Prompt Engineering with Stop Sequences and Message Prefilling

Prompt Engineering and Evaluation

Prompt Evaluation Workflow Key Concepts

Prompt Evaluation Workflow: Key Concepts

Overall Workflow:

I. Purpose of a Grader

I. Purpose of the Code Grader

Improving Model Grader Context via Two-Step Prompt Engineering

I. Core Concepts

Goal of Prompt Engineering:** To improve the quality and outcome of AI responses.

Prompt Engineering: The Concept of Specificity

Purpose of XML Tags in Prompt Engineering:** XML tags are used to provide explicit structure within a prompt, helping the AI (like Claude) better understand how different pieces of content are grouped and what their function is.

Core Concept:** Providing examples within the prompt to guide the AI's desired output.

Prompt Engineering Concepts: Improving Output Reliability

Tool Use in Claude: Study Notes

Goal: Implementing Future Reminder Functionality for Claude

Tool Implementation Concepts

Core Concepts

1. Tool Integration Process

I. Executing the Requested Tool (Step 4)

I. Core Concept: Sequential Tool Use

Goal:** The system manages a continuous loop (`run_conversation`) to allow Claude to interact with external tools until a final, non-tool-based response is generated.

Key Concepts: Integrating Custom Tools with Claude

The Challenge of Parallel Tool Calls

Structured Output using Tools in Claude

I. Overview of the Text Editor Tool

Web Search Tool Study Notes

Definition and Purpose

Goal:** To generate answers based on external, proprietary knowledge.

RAG Pipeline & Semantic Search

RAG Pipeline: Detailed Process Study Notes

RAG Workflow Implementation using a Vector Database

RAG Pipeline Improvement: Combining Semantic and Lexical Search

1. The Need for a Retriever

Topic: Improving Retrieval Accuracy with Re-ranking

Contextual Retrieval for RAG Pipelines

Definition and Purpose

Core Concepts:

PDF Processing Capability:** Clod can read and extract content directly from PDF files, not just images.

Key Concepts: AI Citations Feature

Prompt Caching: Key Concepts

Prompt Caching Key Concepts

I. Core Concepts

Model Context Protocol (MCP) Study Notes

Model Context Protocol (MCP) Study Notes

Project Goal & Purpose

MCP Server Development Study Notes

MCP Server Debugging and Testing

I. Core Concepts & Purpose

Goal:** Implement a feature allowing users to mention documents (e.g., `@report.pdf`) to automatically fetch and insert the document's content into the prompt sent to the LLM (Claude).

MCP Client Functionality:** The MCP client requires a dedicated function (`read resource`) to allow the application to request and retrieve specific resources defined on the MCP server.

MCP Server Prompts: Key Concepts

MCP Client and Prompt Management

Server Primitives: Tools, Resources, and Prompts

Anthropic Applications:** The module focuses on two deployed applications: CloudCode and ComputerUse.

Cloud Code: Key Concepts

Role:** Cloud Code functions as an engineering partner, not merely a code generator.

Cloud Code Functionality:** Cloud Code contains an embedded MCP client, which allows it to connect to external MCP servers to dramatically expand its capabilities.

Core Concept: Parallel Execution

Cloud Code: Monitoring and Automated Error Resolution

Claude Computer Use: Key Concepts

Tool Use and Computer Use Concepts

Workflows and Agents

Workflow: Parallelization for AI Analysis

Definition of Chaining Workflow:

Workflow Concept: Routing Workflow

Agents vs. Workflows

Environment Inspection:** Agents must have a mechanism to inspect their environment after taking an action (and sometimes before) to understand the resulting state.

Workflows vs. Agents: Key Concepts

Anthropic Models and API Usage