← All posts
Course · 30 lessons
AI Advanced: Supercomputing, Kernels & Frontier Systems
How the biggest AI systems really work
A 30-day deep dive for specialists: GPU cluster architecture and interconnects, custom kernels and compilers, distributed training and reasoning models, then agent swarms, MCP and sovereign AI infrastructure.
Starting soon: 30 lessons, one a day.
Week 1 · High-Performance GPU Cluster Architecture & Interconnect Fabrics
- Day 1Beyond Kubernetes: The Architecture of Modern AI SuperclustersComing soon
- Day 2GPU Memory Hierarchy & The Bandwidth WallComing soon
- Day 3Interconnect Topologies: NVLink, NVSwitch, and PCIe Gen 5/6Coming soon
- Day 4AI Network Fabrics: InfiniBand vs. RoCE v2Coming soon
- Day 5NCCL Deep Dive: Optimizing Collective CommunicationsComing soon
- Day 6Diagnosing Physical Cluster DegradationComing soon
- Day 7Slurm vs. Kubernetes for SupercomputingComing soon
Week 2 · Low-Level Performance Engineering & Custom AI Compilers
- Day 8GPU Hardware Architecture from the InsideComing soon
- Day 9Writing Custom GPU Kernels with OpenAI TritonComing soon
- Day 10The Mechanics of FlashAttentionComing soon
- Day 11Modern PyTorch Compilers: TorchDynamo and TorchInductorComing soon
- Day 12Next-Gen Microscaling Precision: FP8, FP4, and MX FormatsComing soon
- Day 13Speculative Decoding & Medusa ArchitectureComing soon
- Day 14Advanced KV-Cache EngineeringComing soon
Week 3 · Distributed Scaling & Advanced Post-Training (RL & Reasoning)
- Day 15The 4D Parallelism ParadigmComing soon
- Day 16DeepSpeed ZeRO & PyTorch FSDP-2Coming soon
- Day 17Mixture of Experts (MoE) Architecture at ScaleComing soon
- Day 18Preference Alignment Beyond SFT: DPO, KTO, and PPOComing soon
- Day 19Reasoning Models & Reinforcement Learning via GRPOComing soon
- Day 20Synthetic Data Flywheels & Automated Curriculum DesignComing soon
- Day 21Model Distillation & Extreme QuantizationComing soon
Week 4 · Autonomous Swarms, Model Context Protocol (MCP) & Enterprise Sovereignty
- Day 22Model Context Protocol (MCP) ArchitectureComing soon
- Day 23Decentralized Multi-Agent SwarmsComing soon
- Day 24Self-Healing and Self-Compiling Code RuntimesComing soon
- Day 25Frontier AI Red-Teaming & Adversarial DefenseComing soon
- Day 26Confidential Computing for AI InfrastructureComing soon
- Day 27Sovereign AI Datacenter ArchitectureComing soon
- Day 28Real-Time Multimodal Streaming PipelinesComing soon
- Day 29AI Infrastructure FinOps & Compute Unit EconomicsComing soon
- Day 30The Post-Transformer HorizonComing soon