Exam Overview & Strategy

# TensorFlow Developer Certificate - Exam Categories

categories = {
    "Category 1: Regression & Classification": {
        "weight": "~25%",
        "skills": [
            "Build dense neural networks (Sequential API)",
            "Regression with single/multiple features",
            "Binary classification with sigmoid output",
            "Multi-class classification with softmax output",
            "Proper data normalization and preprocessing"
        ],
        "key_apis": [
            "tf.keras.Sequential",
            "tf.keras.layers.Dense",
            "model.compile(optimizer, loss, metrics)",
            "model.fit(x, y, epochs, validation_split)"
        ]
    },
    "Category 2: Convolutional Neural Networks": {
        "weight": "~25%",
        "skills": [
            "Image classification with Conv2D/MaxPooling2D",
            "Transfer learning with pre-trained models",
            "Data augmentation with ImageDataGenerator",
            "Binary and multi-class image classification",
            "Handling different image sizes and channels"
        ],
        "key_apis": [
            "tf.keras.layers.Conv2D",
            "tf.keras.layers.MaxPooling2D",
            "tf.keras.applications (MobileNetV2, etc.)",
            "tf.keras.preprocessing.image.ImageDataGenerator"
        ]
    },
    "Category 3: Natural Language Processing": {
        "weight": "~25%",
        "skills": [
            "Text classification and sentiment analysis",
            "Tokenization and sequence padding",
            "Word embeddings (Embedding layer)",
            "LSTM and Bidirectional LSTM networks",
            "Handling variable-length text sequences"
        ],
        "key_apis": [
            "tf.keras.preprocessing.text.Tokenizer",
            "tf.keras.utils.pad_sequences",
            "tf.keras.layers.Embedding",
            "tf.keras.layers.LSTM, Bidirectional"
        ]
    },
    "Category 4: Time Series & Sequences": {
        "weight": "~25%",
        "skills": [
            "Time series forecasting with windowed datasets",
            "Creating training windows from sequential data",
            "RNN/LSTM for sequence prediction",
            "Moving averages and trend analysis",
            "Handling seasonality and noise"
        ],
        "key_apis": [
            "tf.data.Dataset.window()",
            "tf.keras.layers.SimpleRNN",
            "tf.keras.layers.LSTM",
            "tf.keras.callbacks.LearningRateScheduler"
        ]
    }
}

# Study Plan for TensorFlow Developer Certificate

study_plan = {
    "Week 1-2: Foundations": {
        "focus": "Regression & Classification (Category 1)",
        "tasks": [
            "Complete TensorFlow basics (tensors, eager execution)",
            "Build 5+ dense network models from scratch",
            "Practice data normalization and train/val splits",
            "Learn model.compile() options: optimizers, losses, metrics",
            "Complete 3 regression + 3 classification exercises"
        ],
        "resources": [
            "Lesson 2 of this course (with practice models)",
            "TensorFlow official tutorials: basic classification",
            "Coursera: DeepLearning.AI TensorFlow Developer (Course 1)"
        ]
    },
    "Week 3-4: Computer Vision": {
        "focus": "CNNs (Category 2)",
        "tasks": [
            "Build CNN architectures from scratch",
            "Practice transfer learning with MobileNetV2/InceptionV3",
            "Master ImageDataGenerator for augmentation",
            "Handle binary and multi-class image tasks",
            "Complete 5 image classification exercises"
        ],
        "resources": [
            "Lesson 3 of this course (with practice models)",
            "TensorFlow official tutorials: CNN, transfer learning"
        ]
    },
    "Week 5-6: NLP": {
        "focus": "NLP & Sequences (Category 3)",
        "tasks": [
            "Master Tokenizer and pad_sequences",
            "Build LSTM and Bidirectional LSTM models",
            "Practice with IMDB, sarcasm, and BBC datasets",
            "Understand embedding dimensions and vocab size",
            "Complete 5 text classification exercises"
        ],
        "resources": [
            "Lesson 4 of this course (with practice models)",
            "TensorFlow official tutorials: text classification"
        ]
    },
    "Week 7-8: Time Series + Full Practice": {
        "focus": "Time Series (Category 4) + Mock Exam",
        "tasks": [
            "Build windowed datasets for time series",
            "Practice RNN/LSTM for forecasting",
            "Complete Lesson 5 (Time Series practice models)",
            "Complete Lesson 6 (Full Practice Session - 5 exercises)",
            "Take timed mock exam (5 hours, no breaks)",
            "Review Lesson 7 (Exam Day Tips) the night before"
        ]
    }
}

# Prerequisites checklist
prerequisites = {
    "Python": [
        "Comfortable with Python 3.x syntax",
        "NumPy array operations",
        "Basic data manipulation"
    ],
    "TensorFlow / Keras": [
        "tf.keras.Sequential and Functional API",
        "Common layers: Dense, Conv2D, LSTM, Embedding",
        "model.compile(), model.fit(), model.evaluate()",
        "Callbacks: EarlyStopping, ModelCheckpoint",
        "Saving models as .h5 files"
    ],
    "Machine Learning": [
        "Train/validation/test split concept",
        "Overfitting and underfitting",
        "Loss functions: MSE, binary_crossentropy, sparse_categorical_crossentropy",
        "Metrics: accuracy, MAE"
    ],
    "Environment": [
        "PyCharm IDE (basic navigation)",
        "pip install tensorflow",
        "GPU recommended but not required"
    ]
}