Roboflow, a computer vision dev platform, raises $20million

The funds will be used to continue product development and hiring initiatives

Roboflow, a computer vision dev platform, raises $20million

Roboflow, a Des Moines, Iowa-based firm that creates tools for creating computer vision models, has raised $20 million in a series A investment led by Craft Ventures. The company has now raised a total of $22.2 million, with CEO Joseph Nelson stating that the funds will be used to continue product development and hiring initiatives.

According to Grand View Research, the global computer vision sector will be worth $11.32 billion in 2020. However, while the technology has clear commercial implications — computer vision algorithms can be trained to conduct jobs such as detecting gas leaks, counting medications, and monitoring workplaces to impose social separation – companies may find challenges in putting it to use in production. Rebuilding software infrastructure and hiring the appropriate machine learning talent are common challenges for teams.

Roboflow was founded in 2019 by Brad Dwyer and Nelson, who believe that computer vision is a basic technology that may help developers tackle challenges caused by gadgets’ inability to see the world around them. Developers can use the platform to integrate computer vision into their products by uploading photos and videos to train custom or prebuilt computer vision models.

Computer vision is one of those generational technologies that will be adopted by every sector, much like the personal computer or mobile phone. The ability of software to acquire organized information as input is restricted, and that structure is normally provided by a person, according to Nelson. Computer vision, he continued, allows every part of the environment around us to be programmable, resulting in a Cambrian boom of applications. That is why computer vision should be a part of any developer’s arsenal, not just specialist machine learning teams.

Customers can annotate photos while assessing the quality of datasets to prepare them for training using Roboflow. (Most computer vision algorithms need labels to “teach” the algorithm how to classify objects, places, and people.) Developers can use the platform to produce new training data and discover what configurations result in better model performance. Roboflow can then deploy the model to the cloud, edge, or browser, as well as monitor it for edge cases and degradation over time.

According to Roboflow investor Lachy Groom, the promise of Roboflow is reminiscent of Stripe’s early days. He continued, computer vision, like payments, is a key component of infrastructure that needs to be made widely available to developers. Consider how FaceID seeks to seamlessly unlock phones, or how mobile check deposit eliminates the need to queue at a bank while allowing bankers to focus on customer service.

In the burgeoning computer vision development tools market, Roboflow competes with CrowdAI and Chooch, among others. Roboflow, on the other hand, claims to have over 50,000 users, including engineers from half of the Fortune 100, as well as startups, universities, and hardware firms. Pfizer, Walmart, Amgen, and Cardinal are among the clients.