IBM Generative AI Engineering

Description

In this course, you will learn:

• Employers want job-ready capabilities in next-generation AI, machine learning, deep learning, NLP apps, and massive language models in only six months. 
• Create and deploy generative AI apps, agents, and chatbots with Python frameworks such as Flask, SciPy, ScikitLearn, Keras, and PyTorch. 
• Key-generation AI architectures and NLP models, as well as how to use approaches such as prompt engineering, model training, and fine-tuning. 
• Use transformers like BERT and LLMs like GPT for NLP jobs, together with frameworks like RAG and LangChain.

Syllabus:

1. Introduction to Artificial Intelligence (AI)

• Describe AI and its key concepts.
• Showcase how AI applications and use cases can alter our lives and work.
• Recognize AI's potential and impact on businesses and jobs.
• Describe the challenges, limitations, and ethical considerations with AI.

2. Generative AI: Introduction and Applications

• Describe and distinguish generative AI from discriminative AI.
• Describe the capabilities of generative AI and its applications in the actual world.
• Identify the uses of generative AI in various domains and businesses.
• Investigate popular generative AI models and tools for text, code, image, audio, and video generation.

3. Generative AI: Prompt Engineering Basics

• Explain the meaning and importance of prompt engineering in generative AI models.
• Apply best practices for developing prompts and look at examples of effective prompts.
• Use popular prompt engineering tools and approaches to write excellent prompts.
• Investigate commonly used tools for prompt engineering.

4. Python for Data Science, AI & Development

• Learn Python, the most widely used programming language for data science and software development.
• Python programming logic includes variables, data structures, branching, loops, functions, objects, and classes.
• Show that you can use Python libraries like Pandas and Numpy, as well as build code in Jupyter Notebooks, effectively.
• Use APIs and Python modules such as Beautiful Soup to access and scrape data from the web.

5. Developing AI Applications with Python and Flask

• Describe the steps and processes required to create a Python application, including the application development lifecycle.
• Create Python modules, run unit tests, and package applications while following PEP8 coding best practices.
• Explain Flask's features and deploy web apps with the Flask framework.
• Create and deploy an AI-based application on a web server using IBM Watson AI Libraries and Flask.

6. Building Generative AI-Powered Applications with Python

• Explain the fundamental concepts behind generative AI models, AI technologies, and AI platforms like IBM Watson and Hugging Face.
• Integrate and improve big language models (LLMs) using RAG technology to add intelligence to apps and chatbots.
• Use Python modules such as Flask and Gradio to develop web applications that interact with generative AI models.
• Create generative AI-powered applications and chatbots with generative AI models, Python, and other frameworks.

7. Data Analysis with Python

• Write Python code to clean and prepare data for analysis, including managing missing values, formatting, normalizing, and binning data.
• Use libraries like Pandas, NumPy, and Scipy to conduct exploratory data analysis and apply analytical approaches to real-world datasets.
• Use dataframes to manipulate data, summarize it, understand its distribution, perform correlation, and build data pipelines.
• Create and evaluate regression models with machine learning. Use the scikit-learn library to make predictions and decisions.

8. Machine Learning with Python

• In just 6 weeks, you'll master job-ready basic machine learning abilities in Python, including how to use Scikit-learn to construct, test, and evaluate models.
• How to use data preparation strategies and handle bias-variance tradeoffs to improve model performance.
• How to apply key machine learning methods, such as linear regression, decision trees, and SVM, to classification and regression tasks.
• How to assess model performance with metrics, cross-validation, and hyperparameter adjustment to ensure accuracy and reliability.

9. Introduction to Deep Learning & Neural Networks with Keras

• Describe a neural network, a deep learning model, and how they differ. 
• Show an understanding of unsupervised deep learning models like autoencoders and limited Boltzmann machines. 
• Show an understanding of supervised deep learning models like convolutional neural networks and recurrent networks. 
• Create deep learning models and networks with the Keras library.

10. Generative AI and LLMs: Architecture and Data Preparation

• Differentiate between generative AI architectures and models, including RNNs, Transformers, VAEs, GANs, and Diffusion Models.
• Explain how LLMs like GPT, BERT, BART, and T5 are employed in language processing.
• To preprocess raw textual data, use NLP packages like NLTK, spaCy, BertTokenizer, and XLNetTokenizer.
• Make an NLP data loader with PyTorch to do tokenization, numericalization, and padding on text data.

11. Gen AI Foundational Models for NLP & Language Understanding

• Explain how to utilize one-hot encoding, bag-of-words, embedding, and embedding bags to turn words into features.
• Create and use word2vec models for contextual embedding.
• Create and train a simple language model using a neural network.
• Use N-gram and sequence-to-sequence models to classify documents, analyze texts, and transform sequences.

12. Generative AI Language Modeling with Transformers

• Explain the concept of attention mechanisms in transformers, including their role in capturing contextual information.
• Describe language modeling with the decoder-based GPT and encoder-based BERT.
• Implement positional encoding, masking, attention mechanism, document classification, and create LLMs like GPT and BERT.
• Use transformer-based models and PyTorch functions for text classification, language translation, and modeling.

13. Generative AI Engineering and Fine-Tuning Transformers

• Sought-after job-ready skills businesses need for working with transformer-based LLMs for generative AI engineering... in just 1 week.
• How to perform parameter-efficient fine-tuning (PEFT) using LoRA and QLoRA
• How to use pretrained transformers for language tasks and fine-tune them for specific tasks.
• How to load models and their inferences and train models with Hugging Face.

14. Generative AI Advance Fine-Tuning for LLMs

• In-demand gen AI engineering skills in fine-tuning LLMs employers are actively looking for in just 2 weeks
• Instruction-tuning and reward modeling with the Hugging Face, plus LLMs as policies and RLHF
• Direct preference optimization (DPO) with partition function and Hugging Face and how to create an optimal solution to a DPO problem
• How to use proximal policy optimization (PPO) with Hugging Face to create a scoring function and perform dataset tokenization

14. Fundamentals of AI Agents Using RAG and LangChain

• In under 8 hours, organizations can construct AI agents utilizing RAG and LangChain, which requires in-demand job ready skills.
• How to improve prompt design by combining the principles of in-context learning with advanced prompt engineering methodologies.
• LangChain's key principles include tools, components, chat models, chains, and agents.
• How to use RAG, PyTorch, Hugging Face, LLMs, and LangChain technologies for various applications.

15. Project: Generative AI Applications with RAG and LangChain

• Gain hands-on experience by developing your own real-world AI application, which you may discuss in interviews.
• Get hands-on experience loading documents with LangChain and applying text splitting techniques with RAG and LangChain to improve model responsiveness.
• Create and configure a vector database to store document embeddings, then write a retriever to retrieve document segments based on queries.
• Create a simple Gradio interface for model interaction, then use LangChain and an LLM to build a QA bot that can answer questions from loaded documents.