Qdrant, an open-source vector database startup, wants to help AI developers take advantage of unstructured data.

Image credit: Quadrant

For many, the hype train of ChatGPT and generative AI marks the rise of artificial intelligence into the mainstream. But while there is little doubt in the public perception that a massive cataclysm has occurred in the last six months, the growing demand for AI is to power a myriad of emerging use cases. may overtake the infrastructure required for I am trying to deal with it.

Founded in Berlin in 2021, Qdrant targets AI software developers with an open-source vector search engine and database for unstructured data. classified or labeled.

Qdrant today announced that it has raised $7.5 million in seed funding from lead investor Unusual Ventures to take its technology deeper into the commercial arena. This includes him joining 42cap, IBB Ventures, and a handful of his angel backers, including Cloudera co-founder Amr Awadallah. This is in addition to his €2 million ($2.2 million) in pre-seed funding Qdrant raised last year.

unstructured

Beginner-friendly vector databases are intended for storing unstructured data such as images, videos, and text, allowing people (and systems) to search for unlabeled content. This is especially important for scaling large scale language model (LLM) use cases. ) GPT-4 (enhanced ChatGPT), etc.

According to Gartner, unstructured data accounts for 90% of new data generated in enterprises and is growing three times faster than its structured counterpart. At the same time, the majority of AI research and development (R&D) projects never actually make it into production. According to Qdrant CEO and co-founder Andre Zayarni, this is due to a lack of the right tools, and ultimately the transformation from LLM to real-time unstructured data is more It opens up a wealth of opportunities for anyone looking to build useful AI applications.

“Vector databases are a natural extension of[LLM’s]capabilities,” Zayarni explained to TechCrunch. “GPT’s biggest limitation is that it only ‘knows’ about events that happened before the model was trained, whereas vector he is connected to a database, LLM’s virtual ‘memory’ expands in real time.” Able and real-world data. ”

Investors are also paying attention. Just last year, a similar offer for his Qdrant called Pinecone earned him $28 million, but Zayarni sees his Qdrant’s open-source foundation as a key selling point for future customers. .

“Engineers trust open source, and it would be difficult for proprietary software to compete in this market if there were OSS products that offered similar or better offerings,” said Zayarni.

Of course, there are already other open source players, such as Zilliz, a startup that commercialized the Milvus open source vector database and raised $60 million last year. And earlier this month, Chroma secured his $18 million in seed funding to expand its “AI-native” open-source vector database.

Qdrant’s raising of $7.5 million in seed funding speaks to some extent to where investors’ heads are today. A technology that promises to help advance AI and machine learning and extend its capabilities to all developers is clearly an attractive proposition.

Zayarni said Qdrant spent the better part of a month fine-tuning the pitch deck for the seed funding round, receiving the first term sheet two days after submitting the deck, and another two days later. rice field. Or

“We had over 20 VCs interested, and almost all of them wanted to join us as co-investors later. We probably would have received many more offers,” said Zayarni. . “However, his deep experience with OSS (open source software) at Unusual Ventures and its model of being an active operating partner rather than just an investor was very attractive to us, so we decided to work with them. I decided to go to

Today’s funding news comes months after Qdrant launched its managed cloud service. The managed cloud service is designed to empower developers through one-click deployment, automatic version upgrades, backups, and an upcoming database management interface.

With a fresh injection of funding, Qdrant is also working on an enterprise product that can be hosted on-premises or in a private cloud, which is expected to launch later this year, Zayarni said.