Build AI Agents & RAG Apps with LangChain & CrewAI

Learn to download and install the Anaconda Distribution, enabling the latest Python, data science libraries, and IDEs like PyCharm, Spyder, and Jupyter Notebook.

Install and launch Jupyter Notebook via the Anaconda distribution, then create and rename notebooks, edit with markdown, run Python cells, manage checkpoints and kernels, and export to PDF or HTML.

Explore Jupyter notebook markdown features for documentation inside notebooks, including headings, bold and italic text, paragraphs, line breaks, block codes, nested block codes, lists, horizontal rules, links, and images.

download and install Visual Studio Code from the Visual Studio Code website, choose the installer for your operating system, and follow prompts to accept the agreement and add to path.

Enable GPU by installing CUDA toolkit 12.6 and cuDNN, then install PyTorch and TensorFlow via pip, matching CUDA version to PyTorch, or use Google Colab for GPU computing.

Master python basics, including pip package installation and import of libraries like numpy, pandas, and plotting libraries (Plotly, matplotlib, seaborn); learn variables, identifiers, type conversion, and input.

Learn Python control statements and loops, including if/elif/else and while and for with range start, end and step, plus defining and using functions with def, return, and docstrings.

Explore Python list data structure: how to create with square brackets, elements of mixed types, indexing, slicing, and mutably updating lists with append, extend, insert, pop, and remove.

Learn how tuples serve as immutable, mixed-type sequences that support index-based access and slicing, with creation optional parentheses, and enforce non-modifiability of elements to protect confidential data.

Learn to create and manipulate strings in Python, using positive and negative indices and steps to slice, then concatenate, replace, strip, split, join, and zfill.

Explore the Python set data structure as an unordered collection of unique elements. Create sets, add items, and use operations like union, intersection, and difference, including converting to lists.

Master dictionary data structures in Python, learning key-value pairs, the immutable and unique keys, and how to create, access, and manipulate dictionaries with keys(), values(), items(), get(), del, and sorted.

Explore artificial intelligence, machine learning, deep learning, and generative AI, including pre-trained large language models and data lifecycle steps from collection to deployment, plus supervised, unsupervised, and semi-supervised learning.

Explore a brief history of artificial intelligence from the perceptron in 1957 to transformers introduced by Google in 2017, enabling GPT and ChatGPT language models.

Explore transformer architecture, including encoders and decoders, and how attention, positional encoding, and token embeddings power pre-trained models like Bert and GPT.

Explore how transformer attention powers encoder, decoder, and encoder-decoder models across nlp tasks like sentiment classification, named entity recognition, text summarization, and language translation.

Trace the history of foundation models from rule-based and neural networks to the transformer era, highlighting encoder versus decoder roles and the rise of GPT and BERT.

Explore foundation models trained on vast unlabeled data to handle tasks like question answering, sentiment analysis, information extraction, object recognition, and image captioning.

Explore applications of foundation models for automating tasks with reasoning, including customer support, language translation, content generation, image classification, image creation, editing and document extraction, robotics, healthcare, and autonomous vehicles.

Explore the three foundation model types—language models, computer vision models, and multimodal models—and see how text, images, or both power tasks like sentiment analysis, image captioning, and text-to-image generation.

Explore foundation models such as Bert, gpt, jurassic, claude, kahir, stable diffusion, Clip, bloom, Hugging Face, grok, and more, highlighting architectures, data, and capabilities across text, image, and multimodal tasks.

understand language model benchmarks and leaderboards to compare performance across math, coding, and language, explore models and safety options, and learn criteria like accuracy, data quality, size, speed, and ethics.

Explore ethical and responsible AI practices, address biases, and mitigate ten LLM threats from prompt injection to model theft through secure data handling and trusted tooling.

Discover LangChain, a model-agnostic framework that unifies switching between OpenAI, Anthropic, Hugging Face, and more to build end-to-end generative AI apps, including multi-agent workflows with crew AI.

Explore LangChain components to empower chatbots with domain-specific data using Rag retrieval augmented generation, vector databases, tools, and prompt engineering.

Learn how retrieval augmented generation (RAG) uses input preparation, semantic chunking, embeddings, and a vector store to perform similarity search within a RAG pipeline.

Build a rag-based question-answering system with LangChain by preparing domain documents, chunking text, creating embeddings, storing them in a vector database, and running a rag pipeline to answer queries.

Explore vector databases for rag systems, including Pinecone, Chroma, Weaviate, Milvus, and Faiss. Compare features like serverless deployment, real-time vector search, and scalability.

Explore text chunking methods for rag systems, including character-based, word-based, sentence-based, paragraph-based, overlapping, and semantic chunking, to optimize context window usage.

Learn best practices for choosing a chunking method, from character-based to semantic, guided by use case, content structure, and model context, balancing chunking quality with processing speed.

Explore character, word, sentence, paragraph, overlapping, and semantic chunking, then see a character text splitter demo in a rag app using LangChain with 200-character chunks and 50 overlap.

Within the Generative AI Bootcamp, explore the recursive character text splitter from LangChain, using multiple separators, 150 character chunks, and 30 overlap to preserve context in Rag workflows.

Demonstrate the markdown header text splitter for markdown documentation, chunking by headings, using lang chain text splitters and duckling to prepare documents and generate reliable chunks.

Compare token text splitter and markdown header text splitter for reliable chunking, and learn how different chunking methods affect embeddings, vector databases, and Rag application outcomes.

Discover prompt engineering for generative AI by crafting prompts with basic prompt role, task context, few short chain-of-thought, and constrained output prompting, and learn to create an OpenAI API key.

Examine basic prompt engineering techniques through a hands-on demo using LangChain and OpenAI, including rag demos, chunking methods, and a customer messages prompt workflow.

Apply role task context prompts to guide the customer support language model, with a professional role, and clear task, to acknowledge feelings, address concerns, and end with a positive note.

Master few-shot prompt design for customer support, combining role, task, and context prompts with examples. Implement in Python using LangChain and OpenAI to handle angry messages.

Learn how chain of thought prompts enable step-by-step reasoning in customer support, using sentiment analysis, core issues, empathetic responses, and prompt templates to build trust.

Use constrained output prompting with a constraint template to generate a three-paragraph customer response. Adopt a support specialist voice with two empathy sentences, two concrete solutions, and a closing question.

Learn to load CSV, HTML, and PDF documents with Lang chain document loaders, feeding a Rag pipeline with chunking, embeddings, and vector database retrieval for end-to-end generative AI apps.