Complete Gen AI: Basic to Agent AI, RAG, Bedrock, Vertex AI

Master latent space, the compressed, multi-dimensional backbone of generative AI, and see how autoencoders, variational autoencoders, and transformers use it to learn and generate.

Autoencoders compress inputs into a latent space and reconstruct them for compression. Variational autoencoders map inputs to distributions with mean and deviation, enabling sampling to generate new data.

Explore generative adversarial networks (GANs), where a generator and a discriminator engage in adversarial training to produce increasingly realistic data, from images to deepfake video and audio.

Explore the transformer architecture, from input and positional embeddings to encoder-decoder multi-head attention, add & norm, and feedforward networks, illustrated with two word-prediction examples and softmax-based output probabilities.

Explore what NLP is and how it evolved from rule-based to transformer models, uncovering core building blocks like tokenization, stemming, lemmatization, chunking, and embeddings for retrieval and generation.

Explore fundamental NLP concepts such as parts of speech (pos), named entity recognition (NER), chunking, bag of words, tf-idf, and embeddings with practical examples.

Break down text into tokens to enable embedding and similarity search for language models. Explain stemming and lemmatization as reducing words to base or dictionary forms to improve query matching.

Explore how NLP evaluation evolved from rule-based systems to RNN/LSTM and transformers, highlighting how ChatGPT and Gemini deliver rapid, context-aware answers through self-attention.

Walk through setting up Visual Studio Code, Python 3.12, and pip packages, install Neo4j Desktop, and configure free API keys such as Grok for AI workflows.

Build your first llm app in python with grok and openai, wiring api keys, dotenv, and a chat completion flow; then add a simple streamlit UI.

Explore building embeddings and tokenization in Python with NLTK and Gensim, using tf-idf and word2vec, and perform pre-processing with stopwords, punctuation removal, and tokenization.

Explore bag of words and tf-idf vectorization using countvectorizer and tf-idf vectorizer on two simple sentences. Remove stopwords, fit transform, and extract feature names to create a dense numeric matrix.

Learn how prompt engineering drives generative ai by shaping context, role assignment, instruction clarity, and desired output to improve responses across zero-shot, few-shot, and instruction-based techniques.

Explore hands-on prompt engineering with ChatGPT, learning zero-shot and few-shot prompts, role and context setups, and clear instruction and output formats to generate tailored trip plans.

Compare grok UI with ChatGPT, note grok's 500 tokens per second versus GPT-3.5's 40, and use a free grok API key to build prompts.

Learn prompting with Google Gemini, generate trip plans, create accompanying images, and compare results with Google maps, including October Norway itineraries, budgets, and urban cultural details.

Explore how Microsoft Copilot uses prompts to generate a day-by-day trip plan in tabular format and compare its image generation and interface with Gemini.

Learn LangChain as a framework to build your own chatbot by chaining interoperable components like prompts, LM calls, and database connections, enabling scalable AI applications.

Discover how Lang Chain Expression Language (LCL) declaratively defines runnable chains of prompts, LMs, and parsers. Use LCL for asynchronous, parallel execution and streaming, optimizing runtime and readability.

Demonstrates building a Python LangChain program to understand and compare without and with LCL, using env keys, chat templates, chain invocation, and post-processing outputs to a file.

Create a simple LangChain LM app in Python, loading env vars, configuring Grok, and invoking the model; explore retrieval augmented generation and address hallucinations.

Build a simple Streamlit chatbot as the user interface for a language model. Load api keys from an env file and wire in LangChain to enable prompt-based chat.

Explore LangGraph, a land graph framework that implements agentic generative AI with nodes, edges, memory, and tool binding to an LLM.

Explore agentic AI concepts and compare non agent and agent workflows, highlighting iterative, multi agent collaboration that refines drafts through follow up questions and reviewer feedback.