Month 1: Introduction to LLMs and Fundamentals

<aside> 🎯

Goal: To build a foundational understanding of what LLMs are, their history, and their basic architecture.

</aside>

<aside> 📌

• Week 1-2: Begin with the basics.
  • Understand what Large Language Models (LLMs) are.
  • Familiarize yourself with key LLM terminologies.
  • Explore the evolution from language models to LLMs.
  • Look into the history of Natural Language Processing (NLP) and LLMs.
• Week 3-4: Dive into LLM Architectures
  • Study the Transformer architecture.
  • Examine Transformer model design choices.
  • Learn about the Generative Pre-trained Transformer (GPT) architecture.
  • Get an introduction to Large Multimodal Models.
  • Understand the differences between proprietary, open, and open-source language models. </aside>

Month 2: LLMs in Practice and Prompting

<aside> 🎯

Goal: Understand the practical aspects of working with LLMs, including their limitations and how to control their outputs.

</aside>

<aside> 📌

• Week 5-6: Practical aspects of using LLMs.
  • Learn about potential issues like hallucinations and biases in LLMs.
  • Study methods for evaluating LLM performance.
  • Explore how to control LLM outputs.
  • Get introduced to pretraining and fine-tuning of LLMs.
• Week 7-8: Learn about prompting techniques.
  • Understand the basics of prompting and prompt engineering.
  • Explore different prompting techniques.
  • Identify bad prompt practices.
  • Learn tips for effective prompt engineering. </aside>

Month 3: Introduction to LangChain and LlamaIndex

<aside> 🎯

Goal: Introduce two key frameworks for building LLM-powered applications, LangChain and LlamaIndex.

</aside>

<aside> 📌

• Week 9-10: Introduction to LangChain
  • Learn the basics of LangChain.
  • Explore LangChain agents and tools.
  • Understand how to build LLM-powered applications with LangChain.
  • Start building a simple application (e.g., a news article summarizer).
• Week 11-12: Introduction to LlamaIndex
  • Learn about LlamaIndex.
  • Understand the differences between LangChain, LlamaIndex, and OpenAI Assistants. </aside>

Month 4: Prompting and Retrieval-Augmented Generation (RAG)

<aside> 🎯

Goal: Delve into advanced prompting techniques and learn about Retrieval-Augmented Generation (RAG).

</aside>

<aside> 📌

• Week 13-14: Advanced Prompting with LangChain.
  • Learn about LangChain Prompt Templates.
  • Study few-shot prompts and example selectors.
  • Explore output parsers for managing LLM outputs.
  • Improve the news article summarizer application.
  • Learn how to create knowledge graphs from textual data.
• Week 15-16: Begin Retrieval-Augmented Generation
  • Learn what Retrieval-Augmented Generation (RAG) is.
  • Study LangChain’s indexes and retrievers.
  • Understand data ingestion methods.
  • Explore text splitters and their usefulness.
  • Build a simple customer support Q&A chatbot.
  • Learn about embeddings.
  • Understand LangChain Chains.
  • Build a YouTube video summarizer.
  • Explore how to build a voice assistant for your knowledge base.
  • Study techniques for preventing undesirable outputs. </aside>

Month 5: Advanced RAG and Agents

<aside> 🎯

Goal: Explore more advanced RAG techniques and understand how to use LLMs as agents.

</aside>

<aside> 📌

• Week 17-18: Focus on Advanced RAG
  • Understand the differences between prompting, fine-tuning, and RAG.
  • Explore advanced RAG techniques with LlamaIndex.
  • Learn about production-ready RAG solutions with LlamaIndex.
  • Study RAG metrics and evaluation.
  • Learn about LangChain's LangSmith.
• Week 19-20: Focus on Agents
  • Understand what agents are and how large models are used as reasoning engines.
  • Explore AutoGPT and BabyAGI.
  • Study agent simulation projects in LangChain.
  • Build an agent for analysis report creation.
  • Learn how to query and summarize a database with LlamaIndex.
  • Explore building agents with OpenAI Assistants.
  • Learn about LangChain OpenGPT.
  • Work on a project for multimodal financial document analysis from PDFs. </aside>

Month 6: Fine-Tuning and Deployment

<aside> 🎯

Goal: Understand how to fine-tune LLMs and deploy them effectively.

</aside>

<aside> 📌

• Week 21-22: Learn about Fine-Tuning
  • Study techniques for fine-tuning LLMs.
  • Learn about Low-Rank Adaptation (LoRA).
  • Work on a practical example using SFT with LoRA.
  • Study using SFT for financial sentiment.
  • Learn how to fine-tune a Cohere LLM with medical data.
  • Explore reinforcement learning from human feedback (RLHF).
  • Work on improving LLMs with RLHF.
• Week 23-24: Deployment Considerations </aside>