Enhanced Spiking-Convolutional Neural Network Development

I'm seeking a skilled machine learning engineer to design a Spiking-Convolutional Neural Network (SCNN) tailored for image classification from time-series datasets.

create neural network using snntorch module of pytorch only.

Key Specifications:
- The architecture must incorporate two convolutional layers (sconv2d 1 and sconv2d 2) and two fully connected layers (sdense 1 and sdense 2).
- The network should accept an input size of 50x118x1.
- Both convolutional layers are to employ a kernel size of 5x5 with a stride size of 2.
- The first convolutional layer needs to utilize 48 kernels, while the second one will use 24.
- The output from the second convolutional layer (13x30x24) must be flattened and directed to sdense 1 (which should have 256 spiking neurons) and subsequently to sdense 2, yielding 8 output channels.
- Set the batch size to 10, with an initial learning rate of 1e-3 that decays over time.
- The network should be trained with the Adam optimizer.

Datasets:
- The primary datasets for training will be custom image datasets.

Ideal Skills and Experience:
- Proficient in building and training Spiking-Convolutional Neural Networks.
- Experienced with image classification and time-series datasets.
- Familiarity with using ReLU as an activation function.
- Proven track record in implementing and optimizing CNNs.