Udemy
    •  
    •  
    •  
    •  
    •  
    •  
    •  
    •  
Turn what you know into an opportunity and reach millions around the world.
Learn More
Your cart is empty.
Keep shopping
Cisco 300-640 DCAI: AI Data Center Infrastructure Prep
New

Cisco 300-640 DCAI: AI Data Center Infrastructure Prep

Exam-focused prep for Cisco 300-640 DCAI: AI fabrics, RoCEv2, GPU clusters, NDI/NDFC - CCNP DC concentration.
Created byAseem Mankotia
Last updated 9/2026
English

What you'll learn

  • Differentiate training vs inference traffic patterns and their fabric implications
  • Design GPU cluster network fabrics using rail-optimized and fat-tree topologies
  • Configure lossless Ethernet with RoCEv2, PFC, and ECN for AI workloads
  • Apply congestion management and QoS policies for low-latency AI fabrics

Course content

14 sections • 13 lectures
  • AI/ML Workload Fundamentals: Training vs Inference Traffic16:20

Requirements

  • Solid data-center networking fundamentals (350-601 DCCOR-level knowledge recommended)
  • Familiarity with Cisco Nexus 9000 switching platforms

Description

This course contains the use of artificial intelligence.


Built by Aseem Mankotia for data-center, network, and infrastructure engineers, this course delivers exam-focused preparation for Cisco's 300-640 DCAI exam and the Cisco Certified Specialist - Data Center AI Infrastructure credential. You will work through all six official content areas - AI/ML workload and cluster patterns, high-performance AI networking with RoCEv2, PFC, and ECN, AI connectivity and transport models, compute and acceleration, storage and data pipelines, and operations with Nexus Dashboard Insights and Nexus Dashboard Fabric Controller. Every chapter maps directly to a domain from Cisco's published exam outline, with heavily-tested concepts, common exam traps, and scenario-based practice questions called out explicitly so your study time stays targeted.


Aseem Mankotia structures this as a read-and-understand, design-and-reasoning course: no required paid lab, but every topic includes concrete design exercises, topology walkthroughs, and configuration reasoning so the concepts stick. You will learn to evaluate GPU cluster fabric requirements, design lossless Ethernet transport for AI east-west traffic, reason through rail-optimized and fat-tree topology choices, evaluate compute and storage building blocks for training versus inference, and operate AI fabrics using Cisco's assurance and telemetry tooling. The course closes with a full-length practice exam simulation and a time-management strategy session for the 90-minute proctored format.


This credential also counts as a concentration exam toward CCNP Data Center when paired with 350-601 DCCOR, making this course a natural next step for engineers who already hold or are pursuing DCCOR. Because 300-640 DCAI is a newly published exam, always confirm the current official blueprint, section weights, question count, duration, fees, and prerequisites on Cisco's certification page before registering.


AI content disclosure: This course was produced with the assistance of artificial intelligence tools. Lecture narration is AI-voice generated, and lecture scripts, slides, and practice questions were drafted with AI assistance, then reviewed and curated by the instructor for technical accuracy and alignment with the official 300-640 DCAI exam guide.

Who this course is for:

  • Data-center, network, and infrastructure engineers who build and operate the compute, network, and storage behind AI/ML workloads - and those pursuing CCNP Data Center who want the AI-infrastructure concentration. A natural next step after 350-601 DCCOR.