Claude With Amazon Bedrock

Claude Model Families: Key Concepts

AWS Bedrock and Text Generation Study Notes

AWS Bedrock API Request Fundamentals

API Statelessness:** Bedrock and Cloud do not store any messages (user inputs or model responses). They are stateless.

Chatbot Implementation Study Notes

Goal:** To transform a general chat interface into a specialized chatbot that adheres to strict content and behavioral rules.

Prompt Engineering: Controlling AI Output

Input:** Text is fed into the model.

Latency Issue:** Traditional request/response models wait until the entire response is generated before sending it back.

Influencing Claude Output: Advanced Techniques

Study Notes: Controlling LLM Output for Structured Data

Controlling LLM Output with Stop Sequences and Message Prefilling

Prompt Engineering and Evaluation

Prompt Evaluation Workflow Key Concepts

I. Workflow Goal and Constraints

I. Core Workflow

Prompt Evaluation Workflow: Grading Systems

I. Code Grader Functionality

Improving Model Graders via Context Injection

Prompt Engineering Fundamentals

Prompt Engineering Technique: Clarity and Directness

Prompt Engineering: The Concept of Specificity

Purpose of XML Tags:** XML tags are used to provide explicit structure within a prompt, helping the AI (like Claude) understand how content is grouped and what the nature of that content is.

Prompt Engineering: Example-Based Prompting

Prompt Engineering Techniques: Improving Output Reliability

Tool Use in LLMs (Claude)

Goal of the Project:

I. Core Concepts

Tool Inclusion:** When making a request to Claude, include the JSON schema specification that describes how to call your tool.

I. Handling Tool Use Requests

Key Concepts: Integrating Tool Results into Conversation History

Key Concepts for Tool-Using Conversations

Project Goal:** The project aims to implement three distinct tools.

Problem with Standard Tool Use:

Structured Output: Prompt vs. Tools

Structured Data Extraction: Flexible Schema Approach

I. Core Capabilities

Retrieval Augmented Generation (RAG) Study Notes

I. RAG Pipeline Overview

RAG Pipeline: Semantic Search and Text Embeddings

RAG Pipeline Study Notes

Overview

RAG Pipeline Improvement: Hybrid Search

Hybrid Search Architecture & Result Merging

Study Notes: Improving Retrieval Accuracy with Re-ranking

Contextual Retrieval for RAG Pipeline Improvement

Definition and Purpose

Function:** Claude can process images included within a user message.

Reading PDF Documents with Claude

Claude Citations Feature: Key Concepts

Prompt Caching Study Notes

Prompt Caching Key Concepts

I. Core Concepts & Threshold

Model Context Protocol (MCP) Study Notes

Model Context Protocol (MCP) Concepts

Project Overview: CLI Chatbot Implementation

MCP Server Implementation Study Notes

MCP Server Debugging and Testing

I. MCP Client Overview

I. Core Concept & Purpose

MCP Client Resource Reading: Key Concepts

I. Purpose of Prompts in MCP Servers

Client Functionality:

Server Primitives: Tools, Resources, and Prompts

Topic:** Building Agents (a key use case of language models).

Cloud Code: Key Concepts

Understanding Cloud Code

Cloud Code Functionality:** Cloud Code includes an embedded MCP client, which allows it to connect to external MCP servers.

Productivity Gains with Parallel Claude Instances

Cloud Code: Monitoring and Automated Error Resolution

Claude Computer Use: Key Concepts

How Computer Use Works with Clod

Agent Concepts: Key Takeaways