Handwritten Digit Classification with CNN

Job ID: 39294998

Budget: ₹12,500 – ₹37,500 INR

I'm looking for a data scientist with strong experience in machine learning and deep learning, specifically in building and evaluating Convolutional Neural Networks (CNNs) using TensorFlow/Keras.

Key components of the project:

- Preprocessing: The project involves using the MNIST dataset. The required preprocessing steps include normalization, data augmentation, and shuffling the data.
- Model Building: You'll need to build a CNN from scratch to classify the handwritten digits.
- Model Evaluation: The performance of the model should be assessed using various metrics such as accuracy, precision, recall, F1-score, and AUC-ROC.
- Optional Enhancements: If you're able to, hyperparameter tuning would be a beneficial addition. Visualizing misclassified digits to gain deeper insights into the model's behavior would also be valuable.

Your expertise in TensorFlow/Keras and understanding of CNNs will be crucial for the success of this project.