Level 2 · Intermediate Intermediate¶
Goal: move beyond the hello-world pipeline — build real audio and vision applications on microcontrollers, design proper datasets, and shrink models further with pruning, distillation, and quantization-aware training.
Modules¶
- Audio Wake Words & Keyword Spotting
- Image Classification on Microcontrollers
- Edge Impulse End-to-End Workflows
- Data Collection & Dataset Design
- Model Pruning & Knowledge Distillation
- Quantization-Aware Training
- ESP32-CAM Vision Projects
- Person Detection (Visual Wake Words)
- Optimizing Memory & Latency on MCUs
- Project — Voice-Controlled Device
Content status
Level 2 module content is being written — every page below has its scope defined and will be filled in next. Level 1 is complete and is the prerequisite for everything here.