The Complete Course for Generative AI for RPA

Learn to set up python projects in VS Code, create and activate conda environments with Python 3.12, run py files and ipi kernel enabled Jupyter notebooks, and manage kernels.

Explore natural language processing and its day-to-day applications, from virtual assistants and email processing to core tasks like language modeling, text classification, information retrieval, and machine translation.

Apply a Kaizen-inspired MLOps approach to data collection, preparation, model development, and deployment for generative AI. Build automated pipelines with versioning, CI/CD, and monitoring to track data and model performance.

Explore Python tuples: learn to create, access, pack and unpack, and use tuple operations, while understanding immutability, nesting, and practical uses in data structures and function arguments.

Explore real world python list use cases, from a to-do list and inventory management to student grades and user feedback, using append, remove, and in checks.

Learn to perform file operations in Python by reading and writing text and binary files, using with open, handling line by line reads, appending, and seeking to reset the cursor.

Learn how single and multiple inheritance work in Python, using car and Tesla examples to show how parent classes provide attributes and methods to children.

Explore polymorphism in Python's object-oriented programming, focusing on method overriding, interfaces and abstract base classes, with practical shape and animal examples.

Explore encapsulation in Python: bundle data and methods, restrict access with public, protected, and private variables, and use getter and setter methods to manage private state in Python's object-oriented programming.

Explore operator overloading in Python by overriding magic methods like __add__, __sub__, __mul__, __truediv__, and __eq__ within a vector class to customize arithmetic and comparisons.

Learn to implement custom exceptions in Python by creating a generic error and a date of birth exception, then raise and handle them with try-except while validating age in form.

Build an end-to-end machine learning app with streamlit, loading iris data via sklearn, training a random forest classifier, and deploying interactive sliders to predict species.

Build a supervised learning classification using iris data set and a decision tree. Apply standard scaler, perform train-test split, and export the model with joblib for Flask and docker deployment.

Explore text preprocessing with NLTK stemming techniques, including Porter, regex, and Snowball stemmers, their benefits and limitations, and compare to lemmatization in natural language understanding.

Explore NLP in deep learning, from sequential data and neural networks to RNNs, LSTMs, GRUs, and CNNs. Learn encoders, decoders, self-attention, transformers, and how these enable generative AI and LLMs.

Learn how MLOps integrates DevOps principles with the machine learning lifecycle, using version control, automated pipelines, CI/CD, and monitoring to enable collaboration and reliable, scalable model deployment.

Encode categorical features with label and one-hot encoding, scale data, and split into train and test, saving encoders and scaler with pickle to prepare an ANN workflow.

Load trained ann model and pickle files, transform input data with one-hot encoding for geography and label encoding for gender, apply scaling, and predict churn probability.

Build an ANN regression model for salary prediction, detailing data preprocessing, label encoding, one-hot encoding, scaling, train-test split, training with early stopping, evaluation, and streamlit deployment.

Determine the optimal number of hidden layers and neurons in an artificial neural network using grid search, cross validation, and heuristics with keras classifier.

Learn to implement word embedding with keras and tensorflow by converting words to indices, using an embedding layer, and padding sentences to a fixed length for RNN training.

Explore the LSTM RNN architecture, including forget gate, input gate and candidate memory, and the output gate, plus the memory cell that blends short-term and long-term memory.

Explore how a long short-term memory network uses forget, input, and output gates to train on restaurant text, converting words to vectors with word2vec and updating memory.

Analyze variants of LSTM RNN, highlighting forget and input gates, memory cell and candidate memory, peephole connections, and coupled gate designs, and introduce GRU as another recurrent unit.

Implement this project with a GRU variant, replacing LSTM in the layers, train it, and compare accuracy, noting GRU has fewer gates.

Explore the problems with encoder-decoder sequence-to-sequence models, where a single context vector fails on long sentences, lowering blue scores. Learn how attention mechanisms and architecture changes address this.