NVIDIA: Building AI Agents with Multimodal Models (English) - Virtual Classroom

This training is intended for:

• Data Scientists
• AI Engineers
• Deep Learning Specialists
• Robotics Engineers
• Software developers working with neural networks
• Professionals working with sensor or imaging data who want to use AI for analysis and prediction

In de Virtual Classroom wordt de klasruimte digitaal nagebootst en volg je live les via de computer. De lessen zijn klassikaal en interactief, waardoor de les niet heel anders is dan normaal. Je kunt tijdens de les direct vragen stellen aan de docent, werken aan opdrachten en discussiëren met medestudenten. Op deze manier kun je veilig vanuit huis studeren. Het enige dat je nodig hebt is een werkende internetverbinding. Je ontvangt duidelijke instructies over het inloggen en het gebruik en er is technische ondersteuning mochten er problemen zijn.

In this training, you will learn how to develop neural networks that combine different data types (from LiDAR scans to text extracted from documents) to enable better predictions and analyses.

Admission

• Basic knowledge of deep learning concepts
• Experience with a deep learning framework such as TensorFlow, PyTorch, or Keras (this training uses PyTorch)

Objectives

• Practical, project-oriented approach
• Combination of theory, video labs, and hands-on assignments
• Use of PyTorch, OCR tools, vector databases, and multimodal AI frameworks
• Step-by-step guidance for building and orchestrating multimodal models

Program guide

Certificate

Upon completion of this training, you will receive the certificate of participation ‘NVIDIA: Building AI Agents with Multimodal Models (English)’ van Computrain.

Content

• Early and late fusion: combining camera and LiDAR data
• Intermediate fusion: designing architectures for multimodal networks
• Cross-modal projection: adapting language models for visual data (Vision Language Models, VLMs)
• Model orchestration: combining models to answer complex questions (e.g., video analysis with Cosmos Nemotron)
• Assessment: adapting a model to handle different input data

Course schedule

2 dagen

Characteristics

Trainingen

Short program

C-BAIAMM|NVIDIA: Building AI Agents with Multimodal Models (English)

Commercial partners

Brand Owner