Course Component Map

How Read, Build, Explore, and Deploy connect week by week

The ML Systems curriculum has four pillars. This page shows how they integrate across both semesters — so you can see, at a glance, what students do each week.

The Four Pillars

Pillar Resource What Students Do Link
Read Textbook (Vol I or II) Study principles, equations, and case studies mlsysbook.ai
Build TinyTorch Implement framework internals from scratch mlsysbook.ai/tinytorch
Explore Interactive Labs Manipulate simulated hardware, discover tradeoffs mlsysbook.ai/labs
Deploy Hardware Kits Run models on real edge devices mlsysbook.ai/kits

The Theory → Build → Simulation → Reality loop is the core pedagogical cycle. For every major concept (Convolutions, Quantization, Distributed Training), students read the theory, implement it in TinyTorch, explore it in a simulation lab, and (optionally) deploy it on real hardware.

Semester at a Glance

Semester 1: Foundations — Week-by-Week Integration

Week Part Read Build (TinyTorch) Explore (Lab)
1 I Introduction Module 01: Tensor Lab 00
2 I ML Systems Module 01 (cont.) Lab 01
3 I ML Workflow Module 02: Activations Lab 02
4 I Data Engineering Module 02 (cont.) Lab 03
5 II Neural Computation Module 03: Layers Lab 04
6 II NN Architectures Module 04: Losses Lab 05
7 II ML Frameworks Module 05: DataLoader Lab 06
8 II Training Module 06: Autograd Lab 07
9 III Data Selection Module 07: Optimizers Lab 08
10 III Model Compression Module 08: Training Lab 09
11 III HW Acceleration Module 08 (cont.) Lab 10
12 III Benchmarking Catch-up Lab 11
13 IV Model Serving Capstone prep Lab 12
14 IV ML Operations Capstone prep Lab 13
15 IV Responsible Engr. Capstone work Lab 14
16 IV Conclusion AI Olympics Lab 15

Semester 2: Scale — Week-by-Week Integration

Week Part Read Explore (Lab)
1 I Introduction to Scale Lab 01
2 I Compute Infrastructure Lab 02
3 I Network Fabrics Lab 03
4 I Data Storage Lab 04
5 II Distributed Training Lab 05
6 II Collective Communication Lab 06
7 II Fault Tolerance Lab 07
8 II Fleet Orchestration Lab 08
9 III Performance Engineering Lab 09
10 III Inference at Scale Lab 10
11 III Edge Intelligence Lab 11
12 III Ops at Scale Lab 12
13 IV Security & Privacy Lab 13
14 IV Robust AI Lab 14
15 IV Sustainable AI + Responsible AI Lab 15
16 IV Conclusion Lab 16

Hardware Kit Integration Points

Hardware kits are optional but provide powerful “reality checks” at specific moments:

Week (Sem 1) Chapter Hardware Activity Device
4 Data Engineering Sensor data collection Arduino Nano 33 BLE
10 Model Compression Deploy quantized model Seeed XIAO ESP32S3
11 HW Acceleration Profile inference Raspberry Pi + Coral
16 Capstone AI Olympics deployment All three devices
TipNo Hardware? No Problem.

All hardware experiences are replicated in the interactive labs via mlsysim. Labs simulate the exact memory constraints, thermal limits, and latency profiles of real devices.

The Unifying Thread: The Iron Law

Every optimization in both semesters maps to a specific term in the Iron Law:

\[T \approx \frac{D_{vol}}{BW} + \frac{O}{R_{peak} \cdot \eta} + L_{lat}\]

Term Represents Sem 1 Examples Sem 2 Examples
\(D_{vol}\) Data volume Quantization, pruning Gradient compression
\(BW\) Bandwidth Memory hierarchy InfiniBand, all-reduce
\(O\) Operations FLOPs, batch size 3D parallelism
\(R_{peak}\) Peak compute Tensor Cores Multi-node scaling
\(\eta\) Efficiency GPU starvation Pipeline bubbles
\(L_{lat}\) Latency overhead Kernel launch Network latency

Ready to dive into the details? Choose your syllabus: Foundations (Semester 1) | Scale (Semester 2)

Back to top