Live Class 1: Introduction to text preprocessing | What is tokenization and how it works | Breaking text into words or sentences | Introduction to lemmatization | Reducing words to their base form using lemmatization | What are stopwords and why remove them | Identifying and filtering stopwords in text | Using libraries like NLTK or SpaCy for preprocessing | Combining tokenization, lemmatization, and stopwords removal for clean text data

Live Class 2: Introduction to word embeddings | What is TF-IDF (Term Frequency-Inverse Document Frequency) | How TF-IDF captures word importance in a corpus | Introduction to Word2Vec and its two models: Continuous Bag of Words (CBOW) and Skip-gram | Training Word2Vec on text data to create word embeddings | What is GloVe (Global Vectors for Word Representation) | How GloVe differs from Word2Vec and captures global word relationships | Using pre-trained embeddings for NLP tasks | Comparing TF-IDF, Word2Vec, and GloVe for various applications

Project: Sentiment Analysis on Product Reviews

Live Class 1: Introduction to prompt patterns in NLP | Designing effective prompts for different tasks | Few-shot learning and how it reduces data requirements | Understanding few-shot, one-shot, and zero-shot learning | Introduction to OpenAI API | Overview of GPT models and their capabilities | Making API requests to OpenAI for text generation and understanding tasks | Prompt engineering for better results | Fine-tuning models with custom prompts for specific tasks | Using OpenAI API in real-world applications like chatbots, summarization, and translation

Live Class 2: Introduction to function calling with GPT models | Using GPT for structured outputs like JSON, tables, or lists | Designing prompts for structured responses | Calling external functions or APIs from within a prompt | Handling structured data in response formats | Parsing and formatting structured outputs from GPT | Integrating GPT with external systems for dynamic function calls | Use cases of function calling in automation, data extraction, and decision-making tasks | Best practices for optimizing structured output quality

Project: Linkedin Post Generator

Live Class 1: Introduction to LangChain Agent Executor | How the Agent Executor manages complex tasks | Using LangChain to build intelligent agents | Creating and integrating custom tools into LangChain | Leveraging external APIs and services as custom tools | Implementing intermediate memory for maintaining context across tasks | Storing temporary information to guide agent decision-making | Optimizing agent performance with efficient memory management | Example use cases of LangChain in conversational agents and workflow automation

Live Class 2: Introduction to API tools in LangChain | Using built-in API tools for integrating external services | Handling API requests and responses with LangChain | Introduction to file tools for managing document-based inputs | Reading, writing, and processing files in pipelines | Using LangChain to process and analyze file data (CSV, JSON, text) | Error handling strategies in LangChain workflows | Catching and managing exceptions in API and file tool interactions | Debugging and logging in LangChain | Best practices for building robust workflows with error handling

Project: Business Assistant Chatbot

Live Class 1: Introduction to FAISS (Facebook AI Similarity Search) | Using FAISS for efficient similarity search and nearest neighbor search | How FAISS works with high-dimensional data | Introduction to Chroma for scalable vector storage | Using Chroma for real-time, fast retrieval of vector-based data | Overview of LangChain retriever | Integrating LangChain with FAISS and Chroma for document retrieval | How LangChain handles retrieval-augmented generation (RAG) tasks | Benefits of combining FAISS, Chroma, and LangChain for high-performance information retrieval | Use cases in document search, semantic search, and knowledge extraction

Live Class 2: Introduction to the full Retrieval-Augmented Generation (RAG) pipeline | Ingestion of data for RAG systems: collecting and processing raw data | Preprocessing data for embedding generation | Using embeddings to represent data in vector space | Introduction to embeddings: how they capture semantic meaning | Generating embeddings with models like BERT or Sentence Transformers | Storing and indexing embeddings for efficient retrieval | Retrieval process: searching for relevant documents or data based on query

Project: Knowledge Base Chatbot with Vector DB

Live Class 1: Introduction to the attention mechanism | How attention allows models to focus on important parts of the input | Types of attention: hard and soft attention | Self-attention and its role in sequence processing | Introduction to Transformer architecture | Key components of Transformers: Encoder and Decoder | Multi-head attention mechanism | Positional encoding to preserve sequence order | Advantages of Transformers over RNNs and LSTMs | Applications of Transformers in NLP tasks like translation and summarization

Live Class 2: Introduction to Hugging Face Transformers | Overview of BERT and GPT models | Understanding pre-trained models and transfer learning | Fine-tuning BERT for text classification tasks | Preparing the dataset for fine-tuning | Training BERT with custom data using Hugging Face’s Trainer API | Fine-tuning GPT for text generation tasks | Adjusting hyperparameters during fine-tuning | Evaluating model performance on downstream tasks | Using pre-trained Hugging Face models for various NLP applications

Project: Text Classification Using Transformers

Live Class 1: Introduction to chains in NLP workflows | Combining multiple tasks or steps using chains | Using memory in AI models for context retention | Implementing memory to maintain state across interactions | Creating prompt templates for consistent input formatting | Benefits of reusable prompt templates in large-scale applications | How chains and memory work together in task automation | Example use cases of chains, memory, and templates in complex tasks like conversation bots and multi-step reasoning | Optimizing performance with chains and templates

Live Class 2: Introduction to CSV/PDF Q&A pipelines | Extracting data from CSV files for Q&A tasks | Parsing PDF files for text-based question answering | Using OCR for extracting text from scanned PDFs | Building a pipeline to answer questions from structured or unstructured documents | Document indexing for fast retrieval | How to index documents for efficient search and retrieval

Project: Chatbot for Document-Based QA

Live Class 1: Introduction to CrewAI | How CrewAI enables automation and collaboration across tasks | Overview of AutoGen and its role in automating task generation | Using AutoGen for dynamic task creation and assignment | Task delegation strategies for optimizing workflow | Assigning tasks based on skills, priorities, and timelines | Managing task dependencies and ensuring smooth handoffs | Best practices for delegating tasks effectively within automated systems

Live Class 2: Introduction to multi-agent systems | How multiple agents interact and collaborate to solve problems | Example 1: Multi-agent system for recommendation engines | Agents working together to provide personalized suggestions | Example 2: Collaborative robots in manufacturing | Coordinating tasks between robots to improve efficiency | Example 3: Autonomous vehicle fleet management | Multiple agents managing traffic flow and navigation

Project: Career Mate MulltiAgent