Overview
Watch a 33-minute technical talk where Pablo Vega-Behar, Head of AI Implementation at Fitch Group, shares advanced strategies for optimizing RAG (Retrieval-Augmented Generation) pipelines in financial services. Discover innovative approaches to handling high-similarity document sets, including specialized chunking techniques, embedding optimization, and hybrid retrieval methods. Explore real-world applications through detailed demonstrations and learn how Fitch Group's Emerging Tech team enhances retrieval accuracy and generation relevance. Master practical solutions for improving application performance and creating effective development environments that facilitate collaboration between data scientists and ML engineers. Starting with baseline RAG concepts, progress through specific use cases and production system implementation details, before examining current technological limitations. Particularly valuable for advanced data scientists and engineers working with NLP, AI, and machine learning applications in financial services who seek to enhance their RAG systems' performance.
Syllabus
- Introduction
- Baseline RAG
- RAG Use Cases
- Building a Production RAG System
- Demos
- Current Limitations of RAG
Taught by
Open Data Science