Facebook has a new job posting calling for chip designers

Published April 19, 2018

Facebook has posted a job opening looking for an expert in ASIC and FPGA, two custom silicon designs that companies can gear toward specific use cases — particularly in machine learning and artificial intelligence.

There’s been a lot of speculation in the valley as to what Facebook’s interpretation of custom silicon might be, especially as it looks to optimize its machine learning tools — something that CEO Mark Zuckerberg referred to as a potential solution for identifying misinformation on Facebook using AI. The whispers of Facebook’s customized hardware range depending on who you talk to, but generally center around operating on the massive graph Facebook possesses around personal data. While a camera might have a set of data points as a series of pixels, Facebook’s knowledge of you goes well beyond your list of friends and down to minute preferences you have — a set of data so large that it demands a new approach to speed up the process.

Most in the industry speculate that it’s being optimized for Caffe2, an AI infrastructure deployed at Facebook, that would help it tackle those kinds of complex problems. Customized silicon generally tends to be around optimizing inference (the “is that a cat” part of machine learning) or machine training (“this is what a cat is”). On either end, it’s based on speeding up relatively simple math operations based in a field called linear algebra. But we’ve been hearing about this for a bit now, and it seems like Facebook is about to be much more overt about the process.

FPGA is designed to be a more flexible and modular design, which is being championed by Intel as a way to offer the ability to adapt to a changing machine learning-driven landscape. The downside that’s commonly cited when referring to FPGA is that it is a niche piece of hardware that is complex to calibrate and modify, as well as expensive, making it less of a cover-all solution for machine learning projects. ASIC is similarly a customized piece of silicon that a company can gear toward something specific, like mining cryptocurrency.

Facebook’s director of AI research tweeted about the job posting this morning, noting that he previously worked in chip design:

Interested in designing ASIC & FPGA for AI?
Design engineer positions are available at Facebook in Menlo Park.

I used to be a chip designer many moons ago: my engineering diploma was in Electrical… https://t.co/D4l9kLpIlV

— Yann LeCun (@ylecun) April 18, 2018

While the whispers grow louder and louder about Facebook’s potential hardware efforts, this does seem to serve as at least another partial data point that the company is looking to dive deep into custom hardware to deal with its AI problems. That would mostly exist on the server side, though Facebook is looking into other devices like a smart speaker. Given the immense amount of data Facebook has, it would make sense that the company would look into customized hardware rather than use off-the-shelf components like those from Nvidia.

Most of the other large players have found themselves looking into their own customized hardware. Google has its TPU for its own operations, while Amazon is also reportedly working on chips for both training and inference. Apple, too, is reportedly working on its own silicon, which could potentially rip Intel out of its line of computers. Microsoft is also diving into FPGA as a potential approach for machine learning problems.

Still, that it’s looking into ASIC and FPGA does seem to be just that — dipping toes into the water for FPGA and ASIC. Nvidia has a lot of control over the AI space with its GPU technology, which it can optimize for popular AI frameworks like TensorFlow. And there are also a large number of very well-funded startups exploring customized AI hardware, including Cerebras Systems, SambaNova Systems, Mythic, and Graphcore (and that isn’t even getting into the large amount of activity coming out of China). So there are, to be sure, a lot of different interpretations as to what this looks like.

One significant problem Facebook may face is that this job opening may just sit up in perpetuity. Another common criticism of FPGA as a solution is that it is hard to find developers that specialize in FPGA. While these kinds of problems are becoming much more interesting, it’s not clear if this is more of an experiment than Facebook’s full all-in on custom hardware for its operations.

But nonetheless, this seems like more confirmation of Facebook’s custom hardware ambitions, and another piece of validation that Facebook’s data set is becoming so increasingly large that if it hopes to tackle complex AI problems like misinformation, it’s going to have to figure out how to create some kind of specialized hardware to actually deal with it.

Element wants to give identity to the whole world, raising $12M Series A

Published April 11, 2018

Who are you? That’s both an existential question, and also a very practical administrative concern. Today, identity is often exchanged through the use of government ID cards and official paperwork, but what happens when someone loses that paperwork or it is destroyed? Or, as is often the case in many countries around the world, a citizen never received the paperwork to begin with?

Element wants to completely change the way banks, hospitals, and other service providers work with their customers by providing a platform for decentralized biometric identity. The company’s software runs on any mobile device, and using the device’s camera, it can identify a user’s face, palm, and fingerprints to create a verified match. Users have options on which modality they want to use.

Biometric identification is a tough machine learning application, so it shouldn’t be surprising that Element, which was formed in 2012, was co-founded by Adam Perold, a Stanford-educated product designer, and Yann LeCun, a famed machine learning researcher. LeCun was the progenitor of convolution neural nets, which today form one of the foundational theories for deep learning AI. He is now chief science advisor for the company, having taken a role as Director of AI Research at Facebook in New York while continuing his professorship at NYU.

Element is announcing a $12 million Series A round, led by PTB Ventures and GDP Ventures, with David Fields of PTB and On Lee of GDP joining the company’s board of directors. Earlier investors of the company included Pandu Sjahrir, Scott Belsky, Box Group, and Recruit Strategic Partners.

While technologies like Apple’s Touch ID and Face ID systems have popularized biometric identity, neither of these were around when Element got started. The early years of the company were devoted to solving critical technical challenges. Wireless connectivity can be limited in many developing countries, which meant that identities had to be local to the device in order to be useful. That also meant that the platform couldn’t be a cloud infrastructure solution, since identity information had to be processed on the device.

Furthermore, given the quality of hardware available, data had to be extremely compressed to be useful, and the machine learning algorithms couldn’t use too much compute power since a low-powered Android device wouldn’t be able to execute an identity match quickly enough to provide a good user experience.

That’s where LeCun’s deep expertise in neural nets, and particularly in areas like optical character recognition, came in handy. The Element team managed to reduce the amount of data required to store the identity of a single person down to about two kilobytes, according to the company.

The next challenge the company faced in building out its platform was security. Identity data, particularly biometrics, is a major security challenge, but it was exacerbated by the fact that devices would often be shared between users. A single device at a bank, for instance, might service thousands of users, all of which need independent, secured data. The company said that these security challenges have been designed into the core of the system.

Ultimately, the company’s platform lives as an SDK behind the mobile apps of its partners. It provides not only the identity layer itself, but also a secure data infrastructure that allows records such as bank accounts and medical files to be connected to the underlying identity.

Element is targeting the developing world, and Perold tole me he spends more than half of his time traveling to Southeast Asia and Africa building partnerships and doing research on how the company’s technology can improve critical social services. Among the company’s signed partnerships is Telekom Indonesia, which as the service provider for 180 million subscribers, is one of the key connections between people and their identity in that fast-growing economy.

Another partnership formed by the company is with the Global Good Fund, a joint venture between Bill Gates and Intellectual Ventures. That project works to create better biometric identities for newborns and infants, which is critical for health outcomes. The company is working with icddr,b and the Angkor Hospital for Children in Cambodia to build out the program.

In addition to the lead investors, the company received strategic venture capital investments from Bank BCA (via Central Capital Ventura), Bank BRI, Telkom Indonesia (via MDI Ventures), and Maloekoe Ventures.

Correction: The Global Good Fund is a joint venture with Bill Gates, not the Gates Foundation.