metamorworks - stock.adobe.com
Sama, a women-led AI data labeling vendor, said on Thursday it raised $70 million in Series B financing.
Formerly known as Samasource, Sama was founded by Leila Janah in 2008 as a nonprofit. The company changed to a hybrid corporate model in 2019, becoming a for-profit business with the former nonprofit remaining as an investor. Janah died last year.
The vendor, based in San Francisco, with offices in New York and delivery centers in Kenya and Uganda, provides training data to Fortune 2000 companies including Walmart, Nvidia, General Motors and Google. Sama's specialty is in image, video, language and data annotation and validation for machine learning algorithms.
Its investors include CDPQ, First Ascent Ventures, Salesforce Ventures and Vistara Growth.
What $70 million means
While $70 million is a notable number, other data labeling vendors have been raising similar amounts, said Kathleen Walch, an analyst at Cognilytica.
"That Sama raised this much money is great for them but not unheard of in the data labeling space," Walch said. "Quite frankly, they need to start raising money to continue to be a player."
Vendors with similar technology such as Labelbox, CloudFactory and Scale AI have also attracted strong venture funding in recent years. However, Sama's funding success is notable as a woman-led company, because the tech field industry lacks significant gender and other diversity.
"In general, there's a lack of women in AI," Walch said. "For a woman-led company to raise this significant amount of money and be able to be a big competitor in the data labeling space means that women will continue to be taken seriously and we can continue to move forward in this male-dominated industry."
Kathleen WalchAnalyst, Cognilytica
About 40% of Sama's management and leadership are women, and half of its 120 employees are women, according to CEO Wendy Gonzalez. Other women in top leadership include CMO Suzin Wold and senior impact manager Kristen Itani Koue.
"In terms of the development of AI, I think having representation in both the platforms and the data sets that are created are essential to building AI that works for everybody," Gonzalez said.
Gonzalez said Sama has a purposeful hiring model. It tries to hire not only women, but also talent from communities such as Nairobi, Kenya, that are underrepresented in the tech world.
In that sense, the company is mission driven. Its late founder started the company with the idea of providing opportunities to those who didn't have any and trying to lift people out of poverty.
Some members of Sama's workforce live in communities where the household income is less than $2 a day, Gonzalez said. These employees work full-time as data labelers and annotators after going through digital training..
"So, you've got the platform that's being built by a team that's looking for big, broad data sets and then the people who are actually building the data sets are also diverse," she said.
Not only is raising significant venture capital important for a women-led company, doing it during the pandemic is also noteworthy. Women-led startups received just 2.3% of venture capital funding in 2020, according to the Harvard Business Review.
An end-to-end platform
Sama plans to use the Series B money to develop what it says is an end-to-end AI platform.
Gonzalez said Sama's vision includes serving as an indispensable AI data pipeline for data scientists to know what they have in their data sets as well as have the tools they need to train data from the research phase through to building and maintaining the platform.
"Our money is really going to being able to accelerate the development of the end-to-end platform as well as really growing and expanding our global presence," she added.
Machine learning-assisted annotation
In addition to the funding, Sama also introduced its Machine Learning Assisted Annotation MicroModels system. It is available now. Sama declined to provide specific pricing information, saying costs vary on a case-by-case basis.
Sama said the tool powers data set annotation with more efficiency by providing one-click human-in-the-loop validation and speeding up AI development.