Published
04/28/22
share

At GV, we believe that open source platforms will unlock enterprise value in the most collaborative and transparent way. This is why we invested in open source leaders like Cockroach Labs, GitLab, Redpanda, Snorkel AI, Snyk, and Upbound. Today I'm delighted to announce our latest open source investment in deepset — GV's first investment in Germany.

deepset is the creator behind the popular open source library Haystack, a framework that empowers developers to build powerful natural language processing (NLP) applications. Haystack is built in modular fashion, letting users to combine the best technology from other open source projects like Hugging Face's Transformers or Elasticsearch to build search systems that work intelligently over large document collections.

deepset Deepset Co-founders Malte Pietsch, Timo Möller, and Milos Rusic

From my earliest interactions with deepset CEO Milos Rusic and CTO Malte Pietsch, I was struck by their conviction that the company could build a bridge between research and industry using NLP frameworks. As a lawyer by training, I thought back to the many nights I spent poring over documents, wondering when this would be a task that software could do. Remarkable advances in language AI have now taken the potential applications far beyond this example.

Haystack allows developers to rapidly create production-ready pipelines and interact with data through language. Many of the world's largest companies use it to build applications across knowledgement management, enterprise search and virtual assistants. As we spent time learning why these organizations have chosen to build with Haystack, we consistently heard about the tremendous benefits they enjoy across development speed, flexibility and control.

I spoke with Milos about the company's $14M fundraising and today's public launch of deepset Cloud, an enterprise SaaS platform for building language-aware software. Building on the foundations of Haystack, deepset Cloud offers enterprises enhanced abilities around collaboration, customization and managing the entire NLP workflow from ideation to production.

"When we founded deepset in 2018, we felt it was the right time because we had a strong hypothesis around how NLP technology would develop," Milos says. "As open source communities continue to evolve, we will continue to see big jumps in what's possible with NLP at an even faster rate."

Zooming out to the bigger picture, I asked Milos about the market forces driving NLP. "The main drivers are not just the applications but the capabilities," he explains. "There is a lot of data out there, and many efforts to consolidate that data. So the next step is, how can we process that data better, and make it more accessible? NLP can be really powerful here as it allows computers to speak the same language as us and the current state-of-the-art of the technology now truly enables this."

With today's news, deepset's journey enters an exciting new phase. I'm thrilled to welcome Milos, Malte, and the whole deepset team to the GV portfolio.