Tasty Bytes is a fictional global food truck enterprise that has established its presence in 30 cities spanning across 15 countries, boasting a network of 450 trucks offering 15 diverse menu types under various brands. Our mission at Tasty Bytes is committed to improve the Customer Experiences by leveraging the power of AI with Snowflake Cortex. With Snowflake Cortex AI, you gain immediate access to the unsurpassed large language models (LLMs) that have been meticulously trained by esteemed researchers from industry giants like Mistral, Reka, Meta, Snowflake and Google. Most LLMs models today have a very good global performance but fail in specific task-oriented problems. To enhance performance and tackle this challenge, we employ a technique known as Fine-tuning.
With Cortex Fine-Tuning, Snowflake users can harness the power of parameter-efficient fine-tuning (PEFT) to develop custom adaptors for specialized tasks using pre-trained models. This approach avoids the high cost of training a large model from scratch while achieving improved latency and performance compared to prompt engineering or retrieval-augmented generation (RAG) methods.
In this tutorial, we will build a customer support agent that showcases the power of Cortex Fine-Tuning and helps the Tasty Bytes team to respond with a highly accurate automated email to customer tickets, all with minimal resources and time. Fine-tuning has significantly advanced the Tasty Bytes team’s ability to meet the key objective which is nothing but enhancing customer experiences.
For prerequisites, environment setup, step-by-step guide and instructions, please refer to the QuickStart Guide