Simply over a 12 months after launching, Landing AI secured a $57 million spherical of Collection A funding to proceed constructing instruments that allow producers to extra simply and shortly construct and deploy synthetic intelligence techniques.
The corporate, began by former Google and Baidu AI guru Andrew Ng, developed LandingLens, a visible inspection instrument that applies AI and deep studying to search out product defects quicker and extra precisely.
Ng says industries ought to undertake a data-centric method to constructing AI, which gives a extra environment friendly means for producers to show an AI mannequin what to do, utilizing no code/low code capabilities, which embody only a few mouse clicks to construct superior AI fashions in lower than a day.
“We kicked off the data-centric AI motion and are fairly happy that different corporations have began speaking about it,” he advised TechCrunch. “In manufacturing, each manufacturing unit does one thing totally different, so the issue turns into how we will help 10,000 producers construct 10,000 totally different fashions with out having to rent plenty of manpower.”
Andrew Ng, founder and CEO. Picture Credit: Touchdown AI
AI is anticipated to create $13 trillion of realized worth to the world’s financial system by 2030, according to a McKinsey study. Ng says not a lot of that’s realized but, because it stays difficult to construct totally different AI fashions.
He believes Touchdown AI has cracked the code for constructing these fashions, and raised the Collection A spherical after seeing product market match and wished to have the ability to scale the staff to make the product higher.
McRock Capital, an funding agency targeted on the commercial Web of Issues, led the funding and was joined by Perception Companions, Taiwania Capital, Canadian Pension Plan Funding Board (CPP Investments), Intel Capital, Samsung Catalyst Fund, Far Japanese Group’s DRIVE Catalyst, Walsin Lihwa and AI Fund.
Touchdown AI has made strides in constructing its product, however Ng mentioned the corporate remains to be within the early levels of the data-centric AI motion and needs to make extra progress and innovate on some expertise that’s nonetheless lacking.
For instance, whereas beforehand constructing a facial recognition system with 350 million pictures, he discovered that AI expertise invented for that many pictures doesn’t work as effectively to search out defects. A part of the data-centric motion is to develop instruments to assist area specialists by using lower than 50 pictures to supply a clearer instance of what’s a defect.
“We’ve reached a degree the place it’s working and we’re desperate to scale every thing,” Ng mentioned. “We’ve been fascinated for years on the way to crack the recipe and take AI to different industries, and we’re lastly getting there with data-centric AI.”
