Understand the inner workings of neural networks and train them with Keras.

• Introduction to Neural Networks & Deep Learning

• Activation Functions (ReLU, Sigmoid, Tanh)

• Feedforward Neural Network

• Backpropagation & Gradient Descent

• Learning Rate, Schedulers & Optimizers (SGD, Adam, RMSProp)

• Softmax, Cross-Entropy Loss

• Keras Basics: Sequential API & Functional API

• Fully Connected Layer – Forward/Backward Pass

• Regularization – Dropout, Batch Normalization

• Data Preprocessing & Data Augmentation

• Weight Initialization Strategies

• Babysitting Learning: Overfit detection, TensorBoard Monitoring

• Hands-on: MLP on MNIST / Tabular data (e.g. HR Analytics)

Master CNNs, object detection, segmentation, and deployment.

• Basics of Images, Image Preprocessing

• Convolution: 2D Conv, Forward & Backward

• Pooling, Padding, Stride, Transposed Conv

• CNN Architectures: LeNet, AlexNet, VGG, ResNet

• GPU vs CPU for DL

• Transfer Learning: Inception, MobileNet, fine-tuning

• Semantic Segmentation using UNet

• Object Detection – YOLO, SSD, Region Proposal

• Bounding Box Regressor

• Siamese Networks for Similarity Search

• Hands-on:
• Image Classification with CNN
• Object Detection with YOLOv8
• Visual Search with Embeddings

Train text models from scratch and with BERT.

• Introduction to NLP and Use Cases

• Preprocessing: Tokenization, Lemmatization, Stopwords, Normalization

• Feature Extraction: BOW, TF-IDF, N-Grams

• Word Embeddings: Word2Vec, GloVe, Dense Vectors

• POS Tagging, Named Entity Recognition

• RNN, LSTM – Forward Pass and BPTT

• Advanced LSTM Applications + Architectures

• Attention Mechanism + Encoder-Decoder

• Transformers, BERT, Hugging Face Pipelines

• NLP Evaluation Metrics: BLEU, ROUGE

• Hands-on:
• Sentiment Classifier (LSTM or BERT)
• Deploy NLP Model with Streamlit

Take models from notebooks to real-world applications.

• Saving & Loading Models (Pickle, Joblib, Keras)

• Flask vs FastAPI – Serving ML models

• Streamlit/Gradio for Web Apps

• Hosting Models on Hugging Face Spaces, Streamlit Cloud

• MLflow Intro – Model Tracking & Versioning

• Hands-on:
• Deploy CV or NLP model with Streamlit
• Create API using FastAPI