A Swiss big data pioneer in Silicon Valley

Alumni Portraits

While studying for his doctorate at ETH Zurich, Arno Candel learned that there’s a solution to every problem. He currently serves as Chief Architect of H2O, a big data software platform.

Arno Candel
Bild: Khalil Anvar

Silicon Valley is the epicentre of technological innovation, a place of bold dreams and stratospheric careers fuelled by investors who are famously willing to take risks. The birthplace of Google, Facebook, Twitter and LinkedIn, it is a breeding ground for thousands of start-ups each year; some of them change the world, while many others simply disappear. And it’s right here, in San Jose, where one man is revolutionising how we analyse the vast quantities of data we create every single day. Meet Arno Candel, an ETH alumnus, big data expert, Untersiggenthal native – and a beacon of hope for Swiss aspirations in Silicon Valley.

Candel, the “big data all-star”

I meet up with Candel on a sunny Saturday afternoon near his apartment in Santana Row, a small pedestrian zone in San Jose. This is an area of smart restaurants and expensive boutiques, and the boulevard is lined with date palms and patches of bright purple petunias. We take a seat on the terrace of the Left Bank Brasserie, a bistro with Parisian flair. Candel orders a Niçoise salad and a glass of Chardonnay. He tells me he likes coming here, though by the time he leaves work it’s often as late as 10 p.m. before he starts dinner here with his wife, while their son sleeps beside them in his stroller.

Candel has certainly had an eventful year. He became a father last September, shortly after Fortune magazine had chosen him as one of its top 20 Big Data All-Stars, then in April this year he was promoted to Chief Architect at H2O. “What I’m developing here has the potential to change the world,” he says, praising the attraction and excitement of Silicon Valley in a Swiss German Aargau dialect with a few American words thrown in for good measure. He is jointly responsible for developing one of the fastest and most flexible big data analysis platforms for H2O, which is the name of both the company – founded just four years ago – and the software it produces. Analysts argue that data is the new oil, but the only way to transform mountains of unstructured data – including emails, statistics, tweets and images – into valuable information is by using intelligent analysis software. To illustrate his point, Candel gives an example of a customer from the pharmaceutical industry who worked with H2O to accelerate their speed-to-market – not just with limited data samples, but with all of the available data. This produced a significantly more accurate forecast, which – according to the customer – led to seven million dollars of additional revenue. H2O may not have been around for very long, but the company already has some big-name users: Pay Pal uses its system to detect insurance fraud, while Cisco uses it to rank potential buyers. Meanwhile some hospitals are using the algorithms to calculate which rooms to allocate to patients to keep infection risks as low as possible.

Networking with pizza and presentations

Candel pulls his smartphone out of his jeans and shows me pictures of the H2O headquarters in Mountain View, near the Google campus. They reveal a long and completely openplan space, just a third of which is occupied by employees’ desks. The rest is left clear for presentations and meetings. Every two or three days H2O sends out invitations to people who share an interest in the company’s activities. Up to a hundred guests feast on pizza while H2O employees and guest speakers provide insights into their work. “It’s a great way of getting in touch with potential customers,” says Candel.

H2O’s owners are hoping to go public in two or three years. That may seem odd, since anyone can download the software for free and the company only earns money by providing support and services, but things move fast in Silicon Valley! When he joined H2O 18 months ago, Candel was one of just eight employees – today the company employs 38 people. Candel programs virtually around the clock, improving the system and adding new features. Yet he still finds time to attend exhibitions and conferences and give presentations. He hasn’t taken a vacation in the last 18 months. Even at the weekend he typically spends a few hours working. Since he made the switch from research to industry four years ago, his hobbies, which include golf, photography and the violin, have taken a back seat. But that doesn’t bother Candel: he has no doubt that any sacrifices he has made are easily justified by the opportunities H2O offers.

A keen learner from a musical family

The son of a Turkish mother and a Dutch father, Candel was raised in the Swiss municipality of Untersiggenthal near Baden. For many years his parents played the violin and flute in an orchestra that is now known as the Sinfonia Baden; their son played violin in the Siggenthal Youth Orchestra. Candel describes a happy childhood in the countryside. Learning came naturally to him right from the start, and he attained excellent grades throughout high school. “I always wanted to be up there among the best,” he explains.

Candel decided to follow in the footsteps of his father – still his role model – by focusing on scientific and technical subjects. His father spent his entire career working for the Swiss electrical engineering company BBC, now known as ABB. Soon after he arrived at ETH to study physics, Candel discovered an interest in computers. Ralph Eichler, Candel’s professor and later the president of ETH, offered him the chance to run supercomputer simulations at the Paul Scherrer Institute (PSI) while working on his master’s thesis. “That was my debut in the world of big data,” Candel recalls. He subsequently wrote his doctoral thesis on the computer-aided simulation of electron sources for x-ray free-electron lasers, again in collaboration with the PSI and Ralph Eichler. He says that these experiences gave him a lifelong confidence that “there is no such thing as an unsolvable problem.”

The move to Stanford

Shortly before completing his dissertation in 2005, Candel gave a lecture in St. Petersburg. He caught the eye of a researcher from Stanford University in California, who invited the young physicist to tailor his simulations for SLAC’s particle accelerator. Candel made a good impression and was hired as a staff scientist. In 2011 – by which time he had qualified for a green card – he switched to the private sector, initially to another big data company, Skytree, and then to H2O.

Having finished his glass of wine andsalad, Candel orders a double espresso. Before we leave I ask him: can he still imagine making a life for himself in Switzerland? He glances at the passers-by – a series of faces from all over the world – and then nods ruefully towards the blue sky above him: “It would be tough to get used to the greyness of Baden or Zurich after this!” He explains that he would also miss the amenities that Silicon Valley offers ambitious, constantly busy people like him. He uses an app to do his shopping, which arrives at his door two hours after placing the order. And he can eat in San Jose around the clock, thanks to a combination of apps and home delivery services. But the decision on whether Candel and his family will stay in Silicon Valley also comes down to basic economics. He wants his son to have the same kind of childhood he once enjoyed, with a house, garden, nature and good schools. But that’s far from easy in an area where kindergarten places cost up to 3,000 dollars a month, buying a home is a multi-million-dollar investment, and people accept a mortgage rate of four percent or simply pay for the whole thing in cash. That’s why Candel has set himself a time limit: “I’ll give myself another five years. If I make it, then I’ll be able to guarantee my family a good life here. If not, we’ll look for somewhere else where the cost of living is lower.”

COMPANY PROFILE

H2O is a software platform for big data analysis and machine learning. Founded in 2011 in Silicon Valley by two experts in big data, it enables companies to extract key information from large data sets and use the results to make forecasts. Unlike existing big data software products, H2O is open source and can be adapted and scaled to match any system requirements.