Test students' ability to apply model compression and quantization techniques to deploy neural networks on resource-constrained edge devices. Scope includes pruning and structured sparsity, knowledge distillation, weight and activation quantization (post-training and quantization-aware training), mixed-precision, trade-offs among latency, memory, and energy, evaluation metrics, calibration and accuracy-recovery strategies, and practical toolchains/workflows for profiling and deployment (e.g., TFLite, ONNX Runtime, TensorRT). Tasks require selecting methods, estimating resource/performance impact, and outlining implementation and validation steps.