AIC-400 · GPU & AI Compute
Large-Scale Training & Datacenter Architecture
The thousand-GPU conversation: 3D parallelism, reference architectures, TCO and where the hardware is heading.
Who this course is for
Senior engineers and architects designing large-scale AI infrastructure — hundred-to-thousand-GPU training systems and the datacentre decisions around them.
Prerequisites
Course outline
Day 1 — Multi-node training at scale
- Training across hundreds to thousands of GPUs
- 3D parallelism: data + tensor + pipeline
- DeepSpeed ZeRO-Infinity
- NVLink domains and NVSwitch fabrics at scale
- Gradient synchronisation: hierarchical all-reduce, bucket sizes
- Fault-tolerant and elastic training; exascale challenges: power, cooling, reliability
- Trillion-parameter training realities
Day 2 — AI infrastructure architecture
- DGX SuperPOD architecture: compute, storage, networking
- AMD MI300X cluster design
- Reference architectures for AI datacentres
- Rack design: power, cooling, cabling
- Network topology: rail-optimized, dragonfly+
- Multi-vendor integration; TCO analysis
- Firmware lifecycle: BIOS/UEFI, BMC, DPU bundles
Day 3 — Emerging technologies and architecture defence
- CXL 3.0 fabrics and disaggregated memory pooling
- Composable infrastructure: GPU/memory/storage disaggregation
- Serverless GPU (KServe, Knative); WASM runtimes
- Edge AI deployment patterns
- Green AI and sustainability
- Capstone: full architecture review
Hands-on labs
Labs follow the academy model — 35% principles, 20% guided investigation, 45% engineering studio. Every claim you make in a lab is backed by a trace, a counter or a measurement you captured yourself. How we teach
- Lab: model the scaling efficiency of a 3D-parallel training configuration and identify where it breaks
- Lab: design a rack-level layout — power, cooling, cabling, network rails — and cost it with a TCO model
- Lab: evaluate a CXL memory-pooling scenario against a real LLM inference memory problem
- Lab: critique a flawed cluster reference architecture (supplied) and produce a defensible redesign
Capstone project
Design a complete AI datacentre deployment for a stated workload (e.g. training a 70B+ model on a fixed budget): compute fabric, network topology, storage tier, fault-tolerance strategy, firmware/lifecycle plan and TCO — then defend it in a simulated architecture review against cost and reliability challenges.
What you leave with
- A structured method for large-scale training architecture
- Rack-to-fabric design experience with real constraints
- A TCO model you can reuse in procurement discussions
- An evidence-based position on CXL/composable/edge trends
- The design vocabulary of DGX SuperPOD-class systems
How it runs
Every course follows the same model: 35% principles, 20% guided investigation, 45% engineering studio. You leave with working code, raw measurements and an evidence-based report — not a certificate of attendance. Read the methodology or see a full sample lesson.
Material is adapted to your kernel version, hardware and workload before a private delivery. For public cohorts, the environment is provided and configured.
Questions
Who is this course for?
Senior engineers and architects designing large-scale AI infrastructure — hundred-to-thousand-GPU training systems and the datacentre decisions around them. It sits at advanced level within the GPU & AI Compute track.
What do I need to know already?
Specific prerequisites for this course: AIC-210 and AIC-230 (or equivalent production experience); AIC-300; Architecture-level thinking: you have operated real clusters. We confirm levels before the cohort starts and adapt if a group is stronger or weaker than expected.
Can this run privately for my team?
Yes. Any course runs on-site at your offices anywhere, or live online for a distributed team, with labs adapted to your hardware and codebase.
What is the difference between in-person and online?
In person is 3 full days with hardware on your desk, capped at 14. Online is 6 half-day sessions across about two weeks so you can keep working, capped at 20, with remote lab access.
Do you invoice companies?
Yes. Purchase orders are accepted and invoicing is available in USD, EUR, GBP, SAR and CAD.
Upcoming dates
| Dates | Where | Seats | Early bird | Regular | |
|---|---|---|---|---|---|
| 8 Nov – 10 Nov 20263 full days | RiyadhIn person · KAFD Conference Centre | 3 of 14 | SAR 8,100until 9 Oct | ||
| 15 Nov – 17 Nov 20263 full days | Kuwait CityIn person · Al Hamra Tower | 8 of 14 | KWD 670until 16 Oct | ||
| 22 Nov – 24 Nov 20263 full days | MuscatIn person · Knowledge Oasis Muscat | 3 of 14 | OMR 830until 23 Oct | ||
| 22 Nov – 29 Nov 20266 half-days | Gulf bandLive online · 09:00–13:00 GMT+3 | 5 of 20 | US$1,580until 23 Oct | ||
| 23 Nov – 25 Nov 20263 full days | OttawaIn person · Kanata North Tech Park | 8 of 14 | CAD 2,930until 24 Oct | ||
| 30 Nov – 2 Dec 20263 full days | TorontoIn person · MaRS Discovery District | 3 of 14 | CAD 2,930until 31 Oct | ||
| 30 Nov – 7 Dec 20266 half-days | Europe bandLive online · 09:00–13:00 CET | 10 of 20 | US$1,580until 31 Oct | ||
| 7 Dec – 9 Dec 20263 full days | LondonIn person · Shoreditch Works | 8 of 14 | GBP 1,680until 7 Nov | ||
| 7 Dec – 9 Dec 20263 full days | BerlinIn person · Factory Görlitzer Park | 3 of 14 | EUR 1,990until 7 Nov | ||
| 7 Dec – 14 Dec 20266 half-days | Americas bandLive online · 13:00–17:00 ET | 15 of 20 | US$1,580until 7 Nov |
Dates shown for the next few months. If nothing fits, tell us where and when — cohorts are added on demand, and private delivery can be scheduled any week.
More in GPU & AI Compute
AIC-1003 days
Foundations for AI Compute
The architecture, operating system and networking groundwork every GPU systems engineer is assumed to have and often does not.
Practitioner-taught
SAR 6,750Next 25 Oct
AIC-1103 days
GPU Architecture, Memory & Interconnects
How the hardware constrains your workload: SIMT execution, the memory hierarchy and the fabric between GPUs.
Practitioner-taught
SAR 6,750Next 11 Oct
AIC-2004 days
Linux for GPU Systems
What the kernel is doing underneath your training job, and how to tune it. The layer almost nobody teaches.
Practitioner-taught
SAR 10,500Next 15 Nov
AIC-2104 days
RDMA & AI Cluster Networking
Build and debug the fabric distributed training runs on, from queue pairs up to a tuned NCCL all-reduce.
Practitioner-taught
SAR 10,500Next 1 Nov
AIC-2205 days
CUDA & HIP Programming
Write, profile and optimise GPU kernels on both vendors, including the CUDA-to-HIP porting path.
Practitioner-taught
SAR 13,120Next 11 Oct
AIC-2304 days
Distributed Training with PyTorch
From a single-GPU training loop to sharded multi-node training that survives a node failure.
Practitioner-taught
SAR 10,500Next 15 Nov
AIC-3004 days
Containers, Kubernetes & GPU Schedulers
Run a shared GPU cluster multiple teams can actually use: partitioning, scheduling, quotas and isolation.
Practitioner-taught
SAR 10,500Next 18 Oct
AIC-3103 days
Storage & Data Pipelines for AI
Stop starving your GPUs: parallel filesystems, GPUDirect Storage and pipelines built for sustained throughput.
Practitioner-taught
SAR 7,880Next 22 Nov
AIC-3203 days
MLOps & Inference Serving
Get models off a laptop and onto a GPU endpoint that scales, with the pipeline machinery around them.
Practitioner-taught
SAR 7,880Next 1 Nov
AIC-3302 days
AI Infrastructure Security & Observability
Harden a multi-tenant GPU platform and see what it is doing before users report a problem.
Practitioner-taught
SAR 6,000Next 18 Oct