A Melbourne startup founded by two former Amazon employees to improve online search for related objects has raised US$5.2 million (A$8.1m) in a Seed round led by Blackbird.
Marqo is an AI-powered vector search engine. Vector search is a way to find related objects with similar characteristics using machine learning to find semantic relationships.
The raise was also backed by Creator Fund, January Capital, and the cofounders of AI success story Cohere, Ivan Zhang and Aidan Gomez.
The funds will be used to develop a new form of vector search technology that continuously improves based on user engagement, which differentiates it from existing vector databases.
Cofounder Jesse Clark is a former lead scientist at Amazon Robotics AI and scientist at Nasdaq-listed StitchFix, who teamed up with ex-AWS software engineer Tom Hamer. They’ve gone on to recruit talent from the Bezos empire along with Uber and Goldman Sachs.
Hamer said their machine learning models return more accurate search results by understanding content and meaning. Users can search using text, images, or a combination of both.
The technology initially targets end-user search (ecommerce, marketplaces), but has applications in generative AI, analytics and security. The service is currently in closed beta, allows the machine learning models to automatically learn from user engagement and continuously improve the relevance of the vectors.
“Search is in desperate need of modernisation – the vast majority of search experiences are based on legacy keyword search systems, which provide poor results,” Hamer said.
“Customers want search experiences that anticipate their needs, not just match keywords. This is especially important for businesses where the search bar is the core product, such as ecommerce.
“By automatically improving based on user interaction, Marqo provides highly relevant results, and increases customer conversion rate, order value, and revenue.”
Clark said vector search remains challenging to implement especially for applications requiring real-time search,” he said.
“Vector search is rapidly becoming a must-have for generative AI applications. We want anyone to be able to leverage the latest machine learning models, even if they’re not an expert in this field.”
The pair have made the core Marqo code open-source, so it’s free and available for anyone to use, alongside a fully managed, serverless cloud service. The continuous-learning vector search service is expected to be released later this year as a part of Marqo Cloud.
Investor Aidan Gomez, CEO of Cohere, said Marqo is making AI useful to businesses by enabling developers to use the best technology with less effort.
“Vector search is a huge growth area. In addition to being at the core of AI search and recommendation systems – vector search has become a must have component of generative AI,” he said.
Blackbird’s Nick Crocker said most of the data washing around the world is unstructured as text, images or video.
“With generative AI creating more content than ever before people, computers and companies need new ways to search” he said
“Marqo represents the next generation of search – hyper-relevant, AI-powered and based on human understanding.”