Master Large Language Models in 10 weeks

Transform your AI career with our intensive LLM bootcamp. Gain hands-on expertise, learn from industry experts, and build real-world AI applications in just 10 weeks!

AI School

Why Choose AI School?

Are you ready to unlock the transformative power of artificial intelligence? Large Language Models (LLMs) are the technology behind ChatGPT. They are revolutionizing how we interact with information and create software. Our intensive 10-week AISchool bootcamp will equip you with the practical skills and theoretical knowledge to thrive in this exciting field. 

Concise and Focused: Our curriculum distills the essentials of LLM technology into 10 impactful weeks. There is no filler material and no hand-waving. We go deep and understand everything from first principles.

Hands-On Learning: Gain real-world experience through coding exercises, project development, and interactive app creation. After the course, you will be a confident practitioner able to train your own LLMs from scratch.

Industry-Relevant Tools: Master essential Python libraries, transformer models, and frameworks necessary to build and deploy LLMs in the real world

Top notch instruction: Your lead instructor is Dr. Agarwal - a machine learning scientist, university professor and nuclear physicist. She will be assisted by builders from top FAANG companies. You are in good hands. Learn More

Start for FREE and money-back Guarantee: If you dont feel like you got your money’s worth, you can get a full refund.

Classes: 10
Duration: 10 Weeks
Prerequisite: Basic Python
Class: 2 hour instruction + 1 hour group work
Delivery: Online
Support: Office hours + slack community
Start Date (Weekend batch): 16 Nov 2024
Tuition: US $1500

Course Overview

This intensive bootcamp equips you with the essential tools and techniques to harness the power of natural language processing (NLP) and large language models (LLMs). You'll gain hands-on experience in building practical NLP applications, ranging from sentiment analysis and text summarization to advanced chatbot development using cutting-edge AI models. By the end, you'll be well-prepared to leverage LLMs to solve real-world problems and embark on an AI career.


Learning Objectives: By the end of this bootcamp, participants will be able to:

  • Understand the fundamentals of NLP and its applications across diverse domains.

  • Be able to preprocess text data, handle noise, and prepare it for analysis.

  • Represent text using various techniques (BoW, TF-IDF, embeddings).

  • Apply traditional machine learning and deep learning models for sentiment analysis, text summarization, and translation.

  • Work proficiently with transformer models like BERT, BART, and FLAN using the Hugging Face library.

  • Utilize large language models (LLMs) for text generation, question answering, and conversation.

  • Build interactive applications using Streamlit or Gradio.

  • Understand prompt engineering techniques for effective interaction with LLMs.

  • Build RAG applications to leverage external knowledge sources alongside LLMs.

  • Explore the potential of AI agents for automating tasks and creating intelligent systems.

  • Understand ethical considerations in NLP development and deployment.

Week 1: LLM/NLP Basics and Python Primer

Introduction and Overview

  • Introduction to NLP and LLMs: Definition, significance, applications (chatbots, sentiment analysis, translation, summarization, etc.).

  • Python for Text Data: Strings, lists, dictionaries.

    Control flow (if-else, loops), Functions, File I/O (Text, JSON).

  • NLP Libraries: NLTK, spaCy, Pandas (brief overview)

    ; significance and applications.

  • Overview of the bootcamp's goals and topics.

Week 2: Text Preprocessing and Tokenization

Text Cleaning & Normalization

  • Removing unwanted characters, symbols, HTML tags.

  • Lowercasing, handling contractions, emojis, numbers.

Stopword Removal, Stemming, and Lemmatization

  • Concepts, purpose, and usage (NLTK).

  • Brief introduction to regular expressions.

Tokenization & Sentence Segmentation

  • Word vs. sentence tokenization, whitespace, rule-based tokenization (NLTK, spaCy).

  • Brief overview of subword tokenization (BPE, WordPiece)

Week 3: Text Representation and NLP

Text Representation

  • Bag-of-Words (BoW): Concept, document-term matrix, limitations.

  • TF-IDF: Weighing word importance, calculation, advantages over BoW.

  • N-grams: Capturing context with word sequences

Word Embeddings

  • Limitations of one-shot encoding.

  • Word2Vec (skip-gram, CBOW), GloVe, and contextual embeddings (BERT).

  • Capturing semantic relationships between words.

  • Hands-on: Calculating word similarities with GloVe.

Project

  • Introduction to Sentiment Analysis: Definition, importance, overview of traditional classifier

  • Hands on coding example

Week 4: Deep Learning and Transformers

Limitations of Traditional NLP

  • Loss of context, complexity challenges.

  • The Rise of Deep Learning: Neural networks and hierarchical representations.

The Transformer Architecture

  • Attention mechanisms (self-attention, multi-head), positional encoding.

  • Encoder-decoder structure.

  • Advantages over RNNs and CNNs.

  • Key Transformer Models: BERT, BART, FLAN (overview)

Week 5: Advanced NLP with Transformers

HuggingFace Transformers Library

  • Introduction and usage.

Evaluating Performance

  • Sentiment Analysis with Transformers: Hands-on coding, fine-tuning on IMDB or twitter  data.

  • Text Summarization: Extractive vs. abstractive, hands-on with pre-trained models (BART, T5).

  • Text Translation: Machine translation overview, hands-on with pre-trained models.

  • Named Entity Recognition (NER): Introduction, applications, hands-on with pre-trained models.

  • Question Answering systems.

Week 6: Large Language Models - Deep Dive

  • Introduction to LLMs: Definition, characteristics, evolution (GPT to GPT-4).

  • Types of LLMs: Encoder-only, decoder-only, encoder-decoder.

  • Use Cases: Chatbots, content generation, translation, summarization, etc.

  • Compute Requirements: GPUs/TPUs, inference on less powerful hardware.

  • LLM Fine-tuning and Instruction Fine-tuning: Concepts and benefits.

  • How to choose the best model: Hugging Face LLM Leaderboard

  • Loading LLMs: Using Hugging Face Transformers, smaller models (Alpaca, Llama-cpp).

  • Basic LLM Querying: Prompt engineering basics, text generation, question answering.

Week 7: Interacting with LLMs and Langchain

Introduction to LangChain 

  • Framework for LLM-powered apps, key components (models, prompts, chains, agents, memory).

  • Document and Sentence Embeddings: Brief overview (e.g., Sentence Transformers).

Advanced Prompt Engineering with LangChain

  • Key prompting techniques: instruction-based, few-shot, chain-of-thought, system message.

Langchain RAG Features

  • Document Loaders, Text Splitters

  • Embedding Models / Sentence Transformers, Index databases & libraries

  • Retrievers, Text Searches, Chat history, Output parsers

  • Hands on programming exercises

Week 8: Advanced LLM Applications

Generic / stable frameworks.

  • Document Loaders, Advance Text Splitting

  • Advanced Embedidngs & Indexing databases

  • Advanced Tools & Agents

  • Hands-on coding, building moduler units for application development

Advanced Applications - Hands on

  • Hands-on: Building Advanced Chatbots & RAG applications

  • Hands-on: Tools and Agents Integration in NLP applications

Week 9: Chatbots, AI Tools & Agents

Introduction to AI Agents and usage of Tools: 

  • Concepts and potential use cases.

AI Agents Deep Dive:

  • Components of AI agents: perception, reasoning, action, learning.

  • Role of LLMs as reasoning engines for agents.

Hands-On:

  • Building a chatbot with agent-like capabilities, integrating LangChain and tools.

Week 10: Deployment, ethics, model evaluation

Deployment Options: 

  • Overview of Streamlit and Gradio for creating interactive web interfaces.

Ethical Considerations in NLP:

  • Bias in data and models.

  • Privacy concerns.

  • Misinformation and accountability.

Model Evaluation and Fine-Tuning:

  • Metrics for different NLP tasks (BLEU, ROUGE, accuracy, etc.).

  • Approaches to fine-tuning pre-trained models.