NVIDIA-Certified Associate - Generative AI Multimodal (NCA-GENM)¶
Exam Overview¶
Entry-level certification for developers contributing to multimodal generative AI (text, image and audio): model development, programming and quality assurance. NVIDIA's code for this exam is NCA-GENM; this guide called it "NCA-MMGA" until 2026-10-22.
Exam code: NCA-GENM
Duration: 1 hour
Questions: 50-60 multiple choice
Cost: $125 USD
Level: Associate
Validity: 2 years from issuance; recertify by retaking the exam
Delivery: Online, remotely proctored; needs a Certiverse account
Prerequisites: A basic understanding of generative AI
Passing score: Not published
Exam Blueprint¶
| Topic | Weight | Where in this guide |
|---|---|---|
| Core Machine Learning and AI Knowledge | 20% | 01 Multimodal fundamentals |
| Data Analysis | 10% | Not covered in the notes yet |
| Experimentation | 25% | 02 Image and video generation, 03 Text and speech |
| Multimodal Data | 15% | 01 Multimodal fundamentals |
| Performance Optimization | 10% | 04 NVIDIA tools |
| Software Development | 15% | 04 NVIDIA tools |
| Trustworthy AI | 5% | 04 NVIDIA tools (best practices) |
Blueprint updated 2026-10-22 from NVIDIA's exam page (checked 2026-10-22). The notes below were written in May 2026 against an earlier domain list that NVIDIA no longer publishes, so they do not yet map one-to-one onto these topics; topics marked "not covered" need outside study.
Quick Links¶
- 📖 Official exam page - blueprint, price, registration
- NVIDIA Certification Program - Registration
- NVIDIA NIM - Inference microservices
- NVIDIA Riva - Speech AI
- NVIDIA Build - Model exploration
- NVIDIA Developer Blog - Technical articles
Career Benefits¶
Job Opportunities¶
- Multimodal AI Developer
- AI Application Engineer
- Computer Vision Engineer
- Speech AI Developer
- AI Solutions Architect