Join us on May 16th as we deep dive going beyond RAG to build powerful and performant Q&A products.
Speakers:
Vishal Vaingankar (CTO Enviate)
Enviate has developed a custom RAG architecture for our primary use case of running Q&A over our customers' requirements and codebase. The unique graph-based nature of both code and requirements creates challenges when incorporating traditional RAG. We demonstrate how, for structured datasets such as code, we can apply graph ML techniques to augment and enhance document embeddings, delivering Q&A systems that significantly outperform out-of-the-box approaches or fine-tuning.
Christian Griset (Founder/CEO Elliptic Labs)
Elliptic is specializing in building superior search indices tailored for AI applications. De facto Vanilla RAG approaches have significant and fundamental limitations, suffering from what we call "semantic dissonance". At Elliptic, we're directly tackling these limitations with a multi-tiered indexing toolchain. Our technology not only ensures reliable and diagnostic retrieval, it is engineered to seamlessly link business-specific data with the specific tasks of your AIs. Elliptic isn't just for demos; we're making production-ready AIs a reality.
Pebblebed
Pebblebed is an early stage technical firm for builders, by builders. The founding partners are Pam Vagata, founding engineer at OpenAI; Keith Adams, founder of FAIR; Luke Byrne, former partner at Tapestry; and Tammie Siew, former Sequoia & GGV-backed founder.