Boosting RAG and Search

​Register to receive an email link to the recording of this episode of the Together Learning Series!

Join Aamir Shakir, CEO of Mixedbread, for an in-depth session on boosting retrieval-augmented generation (RAG) and search performance using the new mxbai-rerank-v2 models. These open-source reranking models, released under the Apache 2.0 license, represent the latest advancements in retrieval quality—foundational for effective AI and RAG systems.

​In this talk, Aamir will walk through the performance breakthroughs enabled by reinforcement learning-based training, highlight benchmark-leading results, and share implementation guidance to help you improve search relevance in your own systems.

​What we’ll cover:

• ​Model Overview & Release: Introduction to mxbai-rerank-v2, open-source availability, and Mixedbread’s mission for simplifying AI integration

• ​Training Approach: How reinforcement learning techniques drive performance gains across multilingual text, code, and tool retrieval tasks

• ​Benchmark Results: Detailed analysis of how mxbai-rerank-v2 compares to leading rerankers across a range of retrieval scenarios

• ​System Integration Guidance: Best practices for deploying these models to boost retrieval quality and relevance in your applications

• ​Enterprise Use Cases: Real-world benefits of improved retrieval in enterprise AI systems, including better knowledge access and workflow efficiency

​This session is designed to equip you with the tools and understanding to integrate high-performance retrieval into your AI stack and accelerate the effectiveness of your RAG systems.