Hugging Face, a new technology unicorn that democratizes machine learning

Just over a year ago, Hugging Face raised 40 million in an epidemic to build its open-source machine learning platform. Today, Clement Delangor’s co-founded start-up has taken a hit: it has raised $ 100 million in Series C for 2 billion and entered the valuable circle of French tech unicorns.

Existing investors are back in the fold, such as Lux Capital, Lee Fixel’s addition fund (investors in Stripe and Spotify), Beatworks, basketball player Kevin Durant, Olivier Pomel, Datadog co-founder and CEO, Florian Duet. Even Thibault Elzière from Upflow and eFounders. They were joined by two well-known venture capital funds, Sequoia Capital and Kotu Management. “We decided to raise with Sequoia because they are the most recognized and invested in Github, with which our platform is compatible (Github is an open source template for code while Hugging Face does the same thing for machine learning, the editor notes.) And has a lot of experience with Coatue’s end-of-life companies. A

AI at the center of business strategy

Hugging Face is one of those lucky start-ups that doesn’t need money, but at the request of the VC, it’s uncomfortable. “We still have a lot of cash because we haven’t spent our last lap yet But the pace of machine learning is such that we’ve taken advantage of it to ramp up and accelerate. “.

Whether it’s for your zoom wallpaper, ordering an Uber, checking your social media feed or searching Google, artificial intelligence is everywhere. And that’s at the heart of business strategy: “There has been a real evolution, all sectors now use algorithms for technological solutions. And we are the reference platform of machine learning. A

For Bloomberg, Pfizer and Samsung customers

For a year, the name of the embracing face-start-up was a reference to a smiling emoticon with both hands open. Embrace – The model has also changed. Although it initially served as an open-source platform for text, it is now available in machine learning applied to images, video and audio. ” We have a wide variety of applications, we have increased the number of models and datasets shared by users. A

The group now has 100,000 models and 10,000 datasets on the platform and has begun monetizing the offer. It offers a freemium model, meaning free access for the majority with paid features or payments for large companies that use it extensively. Of the more than 10,000 companies using the platform, 1,000 have subscribed to a paid model, including Bloomberg, Pfizer and Samsung. “We have examined the possibility of charging access to the platform in large groups.”Joyful Clement Delangu.

Leadership in ethical matters

Thanks to this new funding, co-founders want to continue investing in the community and increase open source use and sharing. He also hired an American-sized Margaret Mitchell, who co-led the Ethical AI team at Google. He specializes in artificial intelligence ethics, bias and diversity in machine learning. “We are going to take the lead on ethical issues, to influence the models to have a more positive impact. Embrace Face is already the largest platform for the ‘model card’, which sets the standard adopted for communication and risk prevention. Another tool, Data measurement is true Allows researchers, for example, to reduce biases to identify biases in their models.

To do this, start-ups are hiring. Its teams have already grown from 30 to 120 a year and will continue to do so, but have not set themselves a measurable goal. “We are realistic and opportunistic. We are looking for people who are influential and want to join us, in all cases.”. Despite many takeover offers, Clement Delanguo assures him that his job is to develop Hogging Face as a major independent player in machine learning. And one day, Nasdaq to keep her two-handed emoji.

Leave a Comment