Dipanjan Sarkar

In this workshop, you will get a comprehensive introduction into the world of training and fine-tuning Large Language Models (LLMs) and Generative AI. Over the course of five modules, you will explore the essentials of LLMs, deep-dive into training and fine-tuning methodologies, and learn about parameter-efficient techniques like Low rank Adaptation (LoRA), Quantized low rank Adaptation (QLoRA) and instruction-based fine-tuning using techniques like Supervised Fine-tuning and Reinforcement Learning with Human Feedback. You will also learn about the advantages, disadvantages and best practices for these techniques. This is not a university lecture. We will follow a hybrid approach where you will learn the concepts behind these techniques and also spend a lot of time with live hands-on sessions, actually training and fine-tuning LLMs using tools like the HuggingFace ecosystem and Unsloth AI.

Additional Points:

• Prerequisites: Basic understanding of Python, NLP and Deep Learning is helpful
• Content Provided: Slides, complete code notebooks, datasets
• Infrastructure: We will be running everything on the cloud on a GPU server, you will be provided with runpod.io credits or you can use Colab for some of the hands-on also
• Important Note: Some of the hands-on, especially SFT and RLHF on LLMs usually take a minimum of 30+ hours so we might pre-run some of them and do a live walkthrough. Rest most of the hands-on everyone can run live along with the instructor in the session
• Session will be interactive so feel free to keep it interesting and ask questions and treat it as a open-ended discussion

Prepare for the ultimate AI showdown at this year’s DataHack Summit! We're bringing together 3 AI experts to tackle one challenging case problem using three of the hottest techniques in the field: Retrieval-Augmented Generation (RAG), Long Context Language Models, Fine-Tuning  and combining these approaches. This is your chance to see these cutting-edge methods go head-to-head and understand their unique strengths and weaknesses.

It offers a rare opportunity to compare and contrast different problem-solving approaches, learn from leading experts, and expand your understanding of how versatile and dynamic the field can be.

Everyone knows how to build RAG systems, but how do you improve them? Retrieval Augmented Generation (RAG) systems have quickly become among the industry's biggest successes for driving Generative AI use cases on custom enterprise data. However, with their success comes a whole list of pain points that can lead to failure or sub-optimal performance in RAG systems. 

This session is inspired by the famous paper “Seven Failure Points When Engineering a Retrieval Augmented Generation System” by Barnett et al., which discusses some of the major challenges and points of failure in RAG Systems. However, clear solutions to these challenges are not mentioned in detail.

This session aims to bridge this gap where we will cover the major challenges and pain points when building real-world RAG systems, which include:

1. Missing Content
2. Missed the Top Ranked Documents
3. Not in Context 
4. Not Extracted
5. Wrong Format
6. Incorrect Specificity
7. Incomplete

Besides discussing the challenges, we will also discuss practical solutions of how we could address these challenges using the latest and best techniques, including:

• Better data cleaning and prompting 
• More intelligent chunking 
• Better retrieval strategies like Reranking and Compression
• Effect of embedding models and how can we fine-tune such models
• Output parsers for better response format adherence
• Query transformations
• Latest advancements in RAG systems like GraphRAG, Agentic RAG, CRAG, RAFT, etc
• Can long-context LLMs help?

The overall structure of the talk would involve discussing each challenge, discussing potential solutions, and also showcasing some of these with hands-on code leveraging popular frameworks like LangChain and LlamaIndex.