Agentic AI Mastery

Explore module two of Agentic AI Mastery as you learn to use the Google Gemini API with prompt templates and lm chain, and build a simple summarization app with LangChain.

Load the Gemini LLM using Google generative AI via long chain Google JNI, create Gemini 2.0 plus with temperature 0.7, and prepare it for tasks like question answering and summarization.

Explore how a prompt template provides a smart, reusable structure to format prompts with variables, using placeholders and the format method for agentic ai mastery topics like retrieval augmented generation.

Learn how a language model and a prompt template form a reusable lm chain pipeline for tasks like question answering and chatting, with memory enabling context and personalized conversations.

Learn to build multi-step workflows with simple sequential chain and sequential chain, including a real-world two-step keyword extractor and blog title generator using prompts and a language model.

Explore how sequential chain enables multi-input, multi-output AI workflows with named variables, providing greater flexibility and traceability as outputs feed into subsequent steps.

Explore memory in language change by implementing a conversational buffer memory that stores history for context-aware dialogue in language models, using LangChain and Google Gemini API.

Explore how conversation summary memory preserves context in long dialogues by summarizing history instead of storing full chat, enabling efficient memory management.

Explore tools and agents in LangChain, learn how agents act with tools, and build a Google Gemini API tool by installing LangChain, setting an API key, and loading Gemini LM.

Learn how external tools such as calculators, search engines, Python, Repl, weather APIs, and calendar empower agents to make dynamic decisions, build math tools, and enable LangChain workflows.

Create an agent using a math tool with a Gemini model, configure a zero short description agent type in the React framework, and enable verbose reasoning to solve math tasks.

Explore vector databases, document embeddings, and context-aware retrieval via Rag, and learn to generate embeddings, set up Google Gemini with LangChain, and build a document Q&A board.

Understand how a Victory Store stores high-dimensional vectors for fast similarity searches, and how Langton supports vector stores like face, chroma db, and pine code for embeddings.

Generate embeddings for document chunks using Google generative AI embeddings and store them in a face vector store with a FIS index for fast similarity search.

Set up a question-answering chain in LangChain by integrating a language model with the Aphis vector store, using a retriever and stop chain type to answer questions from document chunks.

Learn retrieval augmented generation (rag) by building a rag-based qa board with LangChain and Google Gemini, installing packages, configuring a Google API key, and loading a vector store.

Explore retrieval augmented generation (rag) to prepare a knowledge base by building a rag-based q&a board with LangChain and Google Gemini, using a vector store and link chain.

Run code to load a long document, split into 100-character chunks with overlap, and apply auto indexing to the database, creating two chunks ready for the rag chain.

Build a RAG chain that connects a retriever to a language model, using a vector store to fetch documents and return grounded answers with source documents.

Demonstrate how to ask a question using a rag chain to retrieve relevant document chunks from the vector store and generate a grounded answer with the Gemini LM.

Load and configure the Gemini LLM by importing the Google JNI library and Google generative AI class, select the 2.0 plus model, and set temperature to 0.5.

Define a calculator tool using the Lang Dot tools decorator, converting a Python function into a lang chain tool that evaluates string expressions with eval and handles errors.

Learn how to add memory to an agent using the conversation buffer memory from long chain memory, storing chat history to enable the agent to remember prior conversations.

Initialize a conversational agent with memory using a calculator tool and a language model, set as a conventional react description with memory object and verbose logging.

See how agents chat, remember user names, and use a calculator tool to multiply 15 by four, illustrating memory, tool use, and multi-agent thinking in agentic ai mastery.

Discover the sequence specialists and reconstruction artists that power generative ai, from rnn and transformers to gan, vae, and diffusion, plus fine tuning and reinforcement learning.

Explore core generative ai models—sequence specialists, GANs, and reconstruction artists—and learn how fine-tuning and reinforcement guide tool selection for creative tasks.

Explore how generative ai and natural language processing power humanlike conversations, from transformer-based llms to gpt and ChatGPT, and how fine-tuning and rlhf shape safe dialogue, addressing hallucinations and bias.