---
title: [Webinar] Using LLMs To Build Custom Conversational Engines
description: We’ll showcase a concrete example of how to evaluate & fine-tune an LLM on the Retrieval Augmented Generation (RAG) stack.
---
Using LLMs To Build Custom Conversational Engines
Fine-tuning LLMs Successfully
Large Language Models (LLMs) have great potential for speeding up the development of text-based applications, such as chatbots and question-answering interfaces. To use LLMs effectively, fine-tuning is critical. Fine-tuning involves introducing the LLM to new or updated data to customize its performance according to your requirements, including improving accuracy on core topics, enhancing coverage of unique cases, reducing bias and harmful content, and more.
Yet, fine-tuning an LLM is a challenging task! In this webinar, we will cover:
- Why you may need to fine-tune LLMs
- What are the fundamental tasks involved in fine-tuning
- A real-life example of two fine-tuning use cases from our customers
You will also get access to a notebook that enables you to fine-tune an LLM for domain-specific needs.
Join us!
Watch the replay
Stéphanie NguyenProduct Manager at Kili Technology
**Jean Latapy**
Solution Engineer at Kili Technology
More and more products are powered by machine learning. That’s why it's \[capital\] to think about ethics and to make sure \[its\] impact \[is\] positive.“
**Clément Delangue,****Co-founder & CTO @Hugging Face**
in [Why Ethics are important in ML ](https://resources.kili-technology.com/nurture/why-ethics-are-important-machine-learning?hsLang=en)
“Our Data-centric approach to AI has helped us achieve a categorization of our own categories that is accurate in more than ninety-five percents of the videos.”
**Andrea Colonna,**
**VP AI @Jellysmack**
in**[How Jellysmack leverages AI to analyze in real-time billions of videos on social networks and identify market trends](https://resources.kili-technology.com/nurture/jellysmack-leverages-ai-to-analyze-in-real-time-billions-videos-on-social-networks-and-identify-market-trends?hsLang=en)**
“But the real-world experience of those who put them into production shows that (...) it's often the quality of data (...) that makes your AI project succeed or fail.”
Edouard D'Archimbaud
Co-founder & CTO @Kili Technology
in [Data labeling - Best practices for project management & collaboration](https://resources.kili-technology.com/nurture/directory-collaboration-project-management?hsLang=en)
**Labeling Platform for High-quality Data**
One tool to label, find and fix issues, simplify DataOps,
and dramatically accelerate the build of reliable AI
[Book a demo](https://kili-technology.com/book-a-demo)
[mailto:thomas.chazot@kili-technology.com](mailto:thomas.chazot@kili-technology.com)