Your result 

When completing this training, you will be able to:

• Recognize the limitations of LLMs and understand why agent-based systems are needed
• Design structured outputs for function calls and API integrations
• Apply retrieval mechanisms and knowledge graphs for domain knowledge
• Set up and orchestrate multi-agent systems using tools such as LangGraph
• Design resilient and scalable agent systems ready for production

Target audience

This training is intended for: 

• AI developers who want to integrate agent-like intelligence into software
• Machine Learning Engineers who want to develop advanced AI workflows
• Software engineers who want to build agents that intelligently control APIs, databases, and other systems
• Architects who want to design scalable multi-agent systems for enterprise environments
• Product owners who want to implement advanced AI solutions within their products

Approach

This training follows a hands-on approach in which theory is directly linked to practical exercises. Step-by-step, you will build agents that:

• Provide structure to outputs
• Can call external systems
• Enable interaction among multiple agents
• Independently retrieve, process, and act on information
• Are ready for use in real, dynamic environments

You will work with frameworks such as LangGraph, NVIDIA NIM, and supporting tools for monitoring and scalability.

Prerequisites

• Basic knowledge of Large Language Models (LLMs) and AI applications
• Experience with APIs and data structures is recommended
• Some experience with software development or system integration is an advantage

Study materials

• Digital learning environment with access to interactive labs
• Practical assignments for designing, building, and optimizing agents
• Documentation and reference materials on LangGraph, NVIDIA NIM, and orchestration practices
• Assessment to test your knowledge and skills

Certificate 

Upon completion of this training, you will receive the certificate of participation ‘NVIDIA: Building Agentic AI Applications with LLMs (English)’ by Computrain.

Content

• Fundamentals of agent abstraction and LLMs
• Introduction to LLM capabilities and pitfalls, minimal agent building
• Structured output & fulfillment mechanisms
• JSON outputs, domain alignment, stable schemas
• Retrieval mechanisms & environmental tooling
• Building external system interfaces, semantic retrieval over document sets
• Knowledge graphs & document graphs
• Document development toward canonical forms, event loops, state management
• Multi-agent systems & frameworks
• Task distribution across multiple agents, communication and orchestration
• Data flywheels & system hardening
• Using logs and feedback loops to continuously improve agents
• Scaling & productionalization
• Resource management, multitenancy, scalable server architectures
• Final assessment
• Deployment of a multi-tenant agent endpoint and testing through distributed interactions
• Real-time agents & responsible agents
• Multimodal agents, ethical pitfalls, and standards for responsible AI systems