NuMind unveils NLP tool for expediting ML model creation with $3M seed funding

NuMind, the maker of a software tool to create custom NLP models, is announcing its $3 million seed funding, led by Flybridge with participation from big_bets, Carya Ventures, the Pioneer Fund, Velocity Ventures, Sharpstone Capital, and Y Combinator. These funds will aid in progressing NuMind’s proprietary NLP project accelerator, designed to expedite ML model creation for performing a variety of enterprise tasks, as well as increase its accessibility.

"We are living a revolution thanks to large language models such as GPT-4, but tools to take advantage of them are not mature yet. Tackling applications such as content moderation or sentiment analysis still requires ML experts and an intense labeling effort," explained Etienne Bernard, CEO at NuMind. "Our goal at NuMind is to make this revolution accessible to all, starting with language understanding. It's a game changer in terms of speed and efficiency for all the companies needing this type of AI."

NuMind’s project utilizes large language models (LLMs) and its “Interactive AI Development” learning paradigm to reduce manual efforts on behalf of developers, ultimately driving increased productivity, according to the vendor. The company’s initial customers designed NLP models dramatically faster than when employing traditional methods, up to a factor of 10.

"The NuMind founders recognized how LLMs could be leveraged as part of a platform to drive textual understanding in an easy-to-use, accessible manner so developers could speed the creation and deployment of NLP models," said Chip Hazard, co-founder and general partner at Flybridge. "Now organizations can maximize the potential of NLP, building better applications and improving experiences."

LLMs is a highly emphasized component within NuMind’s solution, bringing to life the company’s vision of “Interactive AI Development.” The interactive process features questions, answers, and explanations between a person and machine, which embraces a more natural learning strategy while eradicating extensive labeling efforts and accelerating personalized model creation, according to the vendor.

"Up to now there were two ways to teach computers: either by programming or by giving them many examples,” explained Bernard. “I see LLMs as the key to unlock this third way that we call Interactive AI Development.”

On top of the reduction of labeling for data and its intuitive interface, NuMind’s solution conducts automatic ML to identify the best models. Active learning only labels the most important documents, and live performance reports summarize the strengths and weaknesses of models.

Additionally, NuMind’s models are multilingual, removing the need for translation. Projects created with NuMind’s solution are owned by the creator—not NuMind—so models can be implemented in whichever way required.

To learn more about NuMind’s ML project accelerator solution, please visit https://www.numind.ai/.