We started Superlinked almost 4 years ago to help developers build real-time personalized software and last year we closed our first 6-figure customer contract. Over the years, instead of investing in customer-specific ranking models, we found a way to use open-source and proprietary models to turn complex product listings, content, user behavior and other types of data into vector embeddings. To enable software engineers to query these multi-modal vectors effectively, we have built a mechanism to define and prioritise for relevance, freshness, popularity and other objectives simultaneously. Finally, we incorporated backfills and batch query evaluation into our product to address common data engineering challenges in managing systems into production. Together, we call these systems “the vector computer”.
Today, we are open-sourcing the interactive components of Superlinked with a goal to help the industry run vector-powered systems in production and combine unstructured and structured data for accurate RAG, Search, RecSys and Analytics. Here is how Superlinked fits into the enterprise data stack:
‍
.jpeg)
Starburst powers analytics and ETL workflows for 100s of enterprises, moving and aggregating petabytes of data across a complex data landscape. Our clients rely on this data to answer questions and generate insights, increasingly within Gen AI-powered workflows. Starburst is collaborating with Superlinked to extract and transform complex data into vector embeddings, enabling enterprises to take full advantage of their first party data to differentiate their Generative AI offerings
.png){kind=small}
We at Dataiku have been at the forefront of the AI revolution, deeply involved in navigating and addressing the challenges of data complexity that enterprises face, especially with the rise of GenAI. Vector embeddings are an essential component of our LLM Mesh, providing the LLM with the knowledge it needs for the clearest and most beneficial response. Superlinked and Dataiku are collaborating to enhance traditional vector search by using Superlinked's innovative embedding technology within Dataiku's LLM Mesh, making it possible to implement the whole RAG pattern with no code needed, allowing organizations to harness GenAI safely and at scale.
AI has fundamentally changed the way businesses are interacting with software and data, in both operations and for end-user experiences. Vector databases and embeddings have simplified generative AI and semantic search capabilities into real-time applications, changing the way developers build applications. MongoDB’s partnership with Superlinked aims to make it easier for customers to create and maintain entity-level and sub-entity-level vector embeddings for enterprise retrieval augmented generation and other use cases, including analytics or more standard semantic search and recommendation systems.
Redis powers the largest real-time data applications in the industry with our blazing fast in-memory data structures - so it’s no surprise that our vector search offering leads the adoption of real-time generative AI for enterprise. Redis is partnering with Superlinked to make vector search easier to apply to complex data for use cases like RAG with LLMs and e-commerce recommendation systems.
Vectors have been used by machine learning to encode knowledge for decades, but their adoption outside of the tech giants has been slow. Let’s fix that together:
‍
Lastly, we are hiring for engineering and go-to-market roles in San Francisco, London, Budapest and Tel Aviv - join us to redefine how some of the largest companies and systems represent and access knowledge!
‍
Daniel Svonava,CEO & Co-founder at Superlinked
‍
.svg)
.svg)
