Fine-Tuning Embeddings and Advanced Retrieval

• Audience: Machine learning practitioners
• Hands-on labs: Yes
• Learning path: Advanced Generative AI Engineering with Databricks
• Description: This course offers a deep dive into understanding how embedding models affect retrieval performance. We will begin with a quick recap of embeddings and then explore the motivation behind fine-tuning embeddings. We will go over how to select embedding models and determine the data amount needed for effective fine-tuning of embedding models. We will also identify relevant metrics to evaluate an embedding model during fine-tuning, compare how the fine-tuned model performs upon retrieval, then eventually deploy the model using Databricks Vector Search. Lastly, we conclude with how to evaluate the deployed embedding model’s quality.

This is the first course in the GenAI Engineer Professional pathway. Pre-requisites: Completed GenAI Engineer Associate pathway or equivalent practical knowledge of:

• Understanding of deep learning, including how neural networks work, what loss functions are, etc.
• Basic understanding of how to build an LLM application that involves prompt engineering and retrieval-augmented generation (RAG)