Dev Conference by Neo4j
You only need to register once to attend all sessions.
Session Track: AI
Session Time:
Session description
Join this session for an in-depth exploration of evaluating Retrieval Augmented Generation (RAG) pipelines, which are crucial for enhancing the performance of large language models (LLMs). As RAG becomes increasingly prevalent, understanding its evaluation is vital to identifying and improving weak points. This session delves into the benefits of integrating knowledge graphs for better grounding and explainability, contrasting it with traditional vector-based retrievers. We will discuss the latest LLM-based tools for automated RAG evaluation, and showcase applications using Neo4j-backed RAG pipelines and the RAG Automated Assessment (RAGAS) framework. You will gain valuable insights into advanced RAG evaluation techniques, helping you optimise your own RAG implementations.
Senior Software Engineer, Neo4j
Nathalie Charbel holds a PhD in Computer Science. Her research focused on semantic information retrieval from unstructured document corpora (2018). Transitioning to industry, she joined Nobatek/INEF4 (2019-2022), where she contributed to EU projects, emphasising ontologies and semantic web technologies. She later joined Neo4j, and specialised in query language design (2022-2024). Currently, she's part of Neo4j's generative AI team, leveraging her expertise in combining knowledge graphs and LLMs.
Software Developer and Machine Learning Engineer, Neo4j
Makbule Gulcin Ozsoy is a Software Developer/Machine Learning Engineer, mainly working on recommender systems, ranking and information retrieval.
Machine Learning Engineer, Neo4j
Estelle holds a PhD in Particle Physics but has moved to the fascinating field of Data Science / Machine Learning / AI / you name it almost 10 years ago. She discovered graph databases and graph data science a few years ago and since then have applied her expertise in various domains from logistics to criminal investigations. In the Neo4j GenAI team, she contributes to make the world of (Knowledge) Graph accessible for developers using LLMs.