Sign up to save your podcasts
Or
Share Pydantic AI: Building Production-Grade Agentic AI Systems
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Pydantic AI: Building Production-Grade Agentic AI Systems
பைடான்டிக் ஏஐ: உற்பத்தித் தரத்திலான ஏஜெண்டுகள் சார்ந்த செயற்கை நுண்ணறிவு அமைப்புகளை உருவாக்குதல்
This episode of Exploring Modern AI in Tamil podcast explains how dependencies and tools work together within Pydantic AI agents.
- Focuses on building type-safe production workflows for agent developers.
- Illustrates how to test agent behavior using datasets and evaluation tools.
- Describes how to process message history by summarizing older interactions to save tokens.
- Explains using TypeAdapter to persist and load conversation history for long-term state.
- Provides tips for using type-safe dependency injection to simplify agent unit testing.
- Demonstrates how to generate and validate datasets for evaluating agent performance.
- Offers best practices for organizing agent capabilities and managing toolset complexity.
- Adds details on using Logfire instrumentation to visualize and debug agent tool calls.
- Discusses implementing graph nodes and parallel execution for complex control flow.
- Explains how to serialize and store chat histories in JSON format for persistence.
- Explains how to manage conversation flow using unique identifiers like run and conversation IDs.
- Incorporates human-in-the-loop approvals to manage sensitive agent tool calls safely.
- Includes steps for using custom evaluators when loading datasets from files.
- Follows the five best practices for structuring and naming evaluation datasets effectively.
- Discusses patterns for durable execution to handle long-running agent tasks reliably.
- Explains multi-agent orchestration techniques using Agent2Agent and graph-based control flows.
View all episodes
By Sivakumar Viyalanபைடான்டிக் ஏஐ: உற்பத்தித் தரத்திலான ஏஜெண்டுகள் சார்ந்த செயற்கை நுண்ணறிவு அமைப்புகளை உருவாக்குதல்
This episode of Exploring Modern AI in Tamil podcast explains how dependencies and tools work together within Pydantic AI agents.
- Focuses on building type-safe production workflows for agent developers.
- Illustrates how to test agent behavior using datasets and evaluation tools.
- Describes how to process message history by summarizing older interactions to save tokens.
- Explains using TypeAdapter to persist and load conversation history for long-term state.
- Provides tips for using type-safe dependency injection to simplify agent unit testing.
- Demonstrates how to generate and validate datasets for evaluating agent performance.
- Offers best practices for organizing agent capabilities and managing toolset complexity.
- Adds details on using Logfire instrumentation to visualize and debug agent tool calls.
- Discusses implementing graph nodes and parallel execution for complex control flow.
- Explains how to serialize and store chat histories in JSON format for persistence.
- Explains how to manage conversation flow using unique identifiers like run and conversation IDs.
- Incorporates human-in-the-loop approvals to manage sensitive agent tool calls safely.
- Includes steps for using custom evaluators when loading datasets from files.
- Follows the five best practices for structuring and naming evaluation datasets effectively.
- Discusses patterns for durable execution to handle long-running agent tasks reliably.
- Explains multi-agent orchestration techniques using Agent2Agent and graph-based control flows.