Someone wanted to learn this too, so Grasp built them a personal learning path.
You can keep exploring this one or create a path like this one shaped around you.
Recommended resources
Module 1
Foundations of AI/ML Workloads
Module 2
GPU Architecture and Core Metrics
Module 3
Practical PyTorch for Performance Analysis
Module 4
Diagnosing AI Performance Bottlenecks
Module 5
Hands-On Profiling with PyTorch Profiler and Nsight
Module 6
System-Level Tracing and Low-Overhead Instrumentation
Module 7
Designing a Unified Observability Schema for AI
Module 8
Implementing the Metrics Collection Pipeline
Module 9
Full-Stack Correlation and Visualization
Module 10