Concevez des agents d'IA avec OpenAI AgentKit

Aperçu

This course explores how to design and build intelligent, reasoning-based AI agents using OpenAI tools, combining structured reasoning, function calling, memory, and communication to create dynamic, context-aware systems. Designed for developers and AI enthusiasts who want to go beyond prompt engineering, it demonstrates how modern agent frameworks like AgentKit and the Model Context Protocol (MCP) enable agents to reason, plan, and act autonomously using context, tools, and collaboration.

Through guided lessons and hands-on demonstrations, you’ll learn to set up your development environment, integrate OpenAI’s APIs, and design reasoning-driven workflows that mimic human-like problem solving. You will explore how agents use planning, reflection, and self-correction, implement function calling and tool use, manage short- and long-term memory, and establish agent-to-agent communication for collaborative decision-making.

The course culminates in building a fully functional reasoning agent system with a Streamlit-based UI, integrating prompts, memory, tools, and communication into one cohesive framework. By the end of this course, you will be able to:

- Explain the anatomy of intelligent agents, including reasoning, memory, tools, and context. - Set up the OpenAI API, configure environment variables, and initialize AgentKit for agent development. - Design and implement structured reasoning workflows using prompts and reflection-based logic. - Integrate function calling and tool registration for agents to perform dynamic tasks autonomously. - Add short-term and contextual memory for improved continuity and understanding across sessions. - Build multi-agent communication systems using the Model Context Protocol (MCP). - Develop and deploy an interactive reasoning agent application using Streamlit.

This course is ideal for software developers, data scientists, and AI practitioners who want to build autonomous, reasoning-powered applications using OpenAI’s ecosystem. A working knowledge of Python and basic familiarity with APIs or AI models will be helpful, but no prior experience with agent frameworks is required.

Join us to master the next generation of AI development — and learn how to transform models into intelligent, context-aware agents that think, plan, and communicate like real collaborators!

Programme

• Fondements des agents intelligents
Apprenez les principes fondamentaux des agents intelligents, explorez comment le raisonnement, la mémoire et le contexte façonnent leur comportement, et configurez votre premier environnement de travail avec AgentKit.
• Construction de l'intelligence centrale des agents
Développez les capacités de raisonnement de votre agent en concevant des invites structurées, en intégrant des outils et de la mémoire, et en créant un pipeline de raisonnement cohérent capable de résoudre des problèmes en plusieurs étapes.
• Intégration et déploiement avancés
Donnez vie à votre agent grâce à l'intégration système, l'optimisation et le déploiement—conception de l'architecture, test des performances et lancement d'une interface interactive Streamlit.

---

Enseigné par

Edureka

---

Matières

Artificial Intelligence