Marqo is a vector search platform that utilizes a community-backed embedding inference engine to provide fast image and text retrieval, supporting hundreds of embedding models for seamless integration. It addresses the challenges of relevance and efficiency in search systems by enabling scalable, multimodal search capabilities and comprehensive evaluation of retrieval performance.
Funding
$18.1M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Traditional search systems often struggle with relevance and efficiency, particularly when dealing with unstructured data like images and text, leading to poor search results and hindering user experience. Fine-tuning embedding models and building scalable embedding pipelines require significant infrastructure and expertise.
Solution
Marqo provides an end-to-end vector search platform that simplifies the process of building and deploying AI-powered search applications. The platform offers a community-backed embedding inference engine that supports hundreds of embedding models out-of-the-box, enabling fast image and text retrieval. Marqo automates the generation of tags and collections based on relevance and conversion, and it personalizes results by learning from user interactions. Its proprietary LLM training framework analyzes site data, product catalogs, and business goals to train a personalized AI search engine.
Target Audience
Marqo targets e-commerce businesses, retailers, and other organizations seeking to improve search relevance, optimize results for business outcomes, and deliver personalized experiences.
Features
- End-to-end platform covering training, inference, and storage of embeddings
- Support for multimodal search, combining text and images in a single vector
- Access to state-of-the-art multilingual models for search in over 100 languages
- Horizontally scalable architecture for low-latency searches against multi-terabyte indexes
- Automated generation of tags and collections based on relevance and conversion
- Integration with click-stream, purchase, and event data for personalized search experiences
- Proprietary LLM training framework (GCL) for fine-tuning embedding models
- Support for customized search, allowing boosting of products based on margin, sponsorship status, or other KPIs