Why SWIM.AI is on the edge of tomorrow
PUBLISHED: 23:14 16 August 2018 | UPDATED: 13:06 17 August 2018
Iliffe Media Ltd
Edge computing has come of age this year
To drive – or cycle – along Newmarket Road is to enter a traffic light lottery: when you get a string of greens it’s three minutes from McDonald’s to Elizabeth Way, otherwise it can take eight or nine minutes. But with SWIM.AI’s software you’ll have that three-minute ride every time.
The San Jose-based firm – soon to open an R&D office in Cambridge – offers data insights based on “the edge”. Edge computing takes data from far-flung dedicated Internet of Things sensors or control devices and generates insights in real time – like the fastest way to navigate through a city, or track a part in a supply chain. Using AI and machine learning, the firm’s platform, SWIM.EDX, can whisk you through the thousands of sensors – and send drivers another way if there’s a quicker option.
“If you take one traffic intersection in a small town,” says CTO Simon Cosby, “there are probably 50 sensors involved. That’s just one intersection, so there are tens of thousands of sensors involved in one small town, and hundreds of thousands in a medium-sized town.”
“A town with 60,000 intersections would have three million sensors,” says CEO and co-founder Rusty Cumpston. That means a 10-mile journey could involve an interaction with a million sensors. So if you’re fast enough you can make use of the data and monetise the information platform: it’s game-on time for IoT as edge computing begins to reveal what’s possible.
“The world is a data-hungry beast and there are mountains of data being produced at the edge that no one processes,” says Rusty. “SWIM.AI takes all the data and generates valuable insights. The other thing is that the value of the data expires very quickly – the insight won’t be valuable for a long period of time, so you need to share as many insights as possible. Our innovation is that the insights are processed quickly.”
So a driverless car, or a taxi firm, might pay for the traffic lights data?
“Yes,” says Simon. “We can predict ahead of time when the free-running lights will turn. We publish the API (application data interface) in the cloud for Uber and other autonomous vehicles so they can route their cars through the city and never see a red light.”
“Vehicles are becoming mobile data centres,” says Rusty. “They need information coming at them.”
“Another example is a very large manufacturing facility,” says Simon. “Every part is labelled with a small RFID (radio frequency identification) tag, and all the parts have that so you get millions of sensor readings per day, and the firm wants to know where the parts are. It’s a very simple system to say ‘where are you?’, and watch all the parts come together. All this is learned at the edge and they now have a system able to respond in real-time, all through the supply chain.”
SWIM.AI’s edge-harvesting insights translate across all suppliers and into other sectors too, including “healthcare, retail, manufacturing, oil and gas”, says Rusty.
The firm completed a $10million Series B funding round this summer, with Cambridge Innovation Capital (CIC) the lead investor, plus participation from Silver Creek Ventures and Harris Barton Asset Management and a ‘strategic investment’ from Arm taking the total funding to around $18million.
Arm’s involvement is significant. The firm acquired Stream Technologies in June and Treasure Data earlier this month. Stream “supports physical connectivity across all industry-standard wireless protocols and devices critical for making IoT data accessible”, and Treasure Data is a data analytics specialist. The acquisitions are part of Arm’s aim of having a trillion connected devices by 2035. And SWIM.AI is part of that journey.
“SWIM’s ability to analyse data and apply machine learning at the edge unlocks new IoT use cases by unleashing data that was previously too difficult, slow or expensive to send to the cloud for analysis,” said Damon Civin, principal data scientist at Arm, of the deal. “Their solution complements the Arm Mbed IoT Device Management Platform and our mission of enabling organisations to seamlessly obtain and derive meaning from their IoT data.”
Rusty and Simon dropped in at an Arm conference at Clare College during their visit last week.
“This week has been a stunner,” says Rusty. “With what they’re doing at Arm… I’m just incredibly impressed.”
Another reason for being in town is that SWIM.AI is setting up an R&D facility in Cambridge – “our preferred location is near the station”.
The firm has long-standing links in the city.
“For me this is a second home,” says Simon, who was listed among InfoWorld’s Top 25 CTOs in 2007. “I was a lecturer in the Computer Laboratory from 1994 to 2000, now I’m on the board of Cambridge in America.”
Cambridge in America is the New York-based non-profit organisation which promotes interest in and support for the university in the US.
Simon and Rusty both previously worked at data optimisation firm XenSource, which was founded in 2004 and sold to Citrix in 2007.
“When it was acquired by Citrix I went off to start Sensity Systems, and now we’re all together again,” says Rusty, who sold Sensity to Verizon Communications in 2016.
Rusty’s keen on Cambridge too. The CIC investment began a while ago, and the partnership with Arm could be game-changing for the sector as well as both firms. Was it unusual to find CIC investing overseas?
“Well their investment thesis is expanding,” says the CEO, “and they’re investing in companies outside Cambridge that are doing things in Cambridge. This software and what’s happening in edge intelligence is going to be a growing market opportunity for the next 20 years and they see us an early leader.”
“They’re trying to accelerate machine learning at the edge and there are certain features which involve CPUs and Arm is going to lead the way for affordable sensors, and the application area is both consumer and industrial,” says Simon.
“There’s thousands of applications people have identified for SWIM.EDX,” says Rusty. “We run it as a service, but we can also license it.”
With SWIM.AI’s prowess in edge computing, talking about having all bases covered is no longer just a metaphor.
SWIM.AI has opportunities for specialists including software development engineers, solution architects and a tech lead.