Technology of a small Estonian technology firm Cybernetica has been in the R&D phase for a decade, and the team has only now entered the product packaging phase – But this is not your usual startup story.
Dan Bogdanov, 32, grew up wanting to become a writer. He attended one of the best schools in Pärnu, Estonia, where he spent his days learning French and Latin, and his nights writing short stories (which will get published one day, he swears). But it was not French or Latin that ended up changing his life and shifting decision making paradigms all over the world. It was C++, a programming language, that allowed him to write Sharemind – a solution so unfathomable the Director General of the Estonian Tax Board has called it “the fairytale technology”.
Last month The President of Estonia gave Dan a prize for his outstanding scientific achievements in the IT science for creating the Sharemind technology.
So what is it all about?
Sharemind utilises secure multi-party computing in order to analyse confidential data. Most people don’t care how much data is really gathered about them daily – big data is as ubiquitous as morning coffee. Still, governmental and corporate data mining raises red flags everywhere, preventing the capability of smart decisionmaking.
Digital governments (such as in Estonia and Finland) are sitting on heaps of data, which could be analysed to allow making smarter decisions that are eventually beneficial to the entire society. It is illegal (or, at least, highly unethical) today to perform such analyses, as most of the data gathered is highly confidential and private – think of health records.
Calculating the Future from an Invisible Past
Data is often protected by encryption, which literally means “obscuring the message”. Even if someone would intercept the message, for someone without a key, it would be a scramble of bits and bytes with no apparent coherence. The problem with current technology is that in order to run any sort of statistical analysis on data sets, you kind of have to see what you are analysing. This is why the data usually needs to be decrypted before running an analysis. Obviously there are heavy restrictions towards this due to privacy concerns, which limits the data sets usable by statisticians.
Imagine if the governments could behave like big data companies and find answers not about which cat videos to recommend to you next on YouTube, but where to invest public money, how to improve the health system and which industries to subsidize over the next budget period?
Dan did, and his answer was simple: if decryption breaches privacy, let’s keep the encryption and compute using encrypted data. Like with all things, simplicity is deceiving and no matter how easy anything sounds, the actual solution is highly complex. “Three PhD theses have been written on Sharemind, and I have lost count of the published peer-reviewed articles already. We celebrated Sharemind’s 10th birthday in January with our team at Cybernetica and honestly, we are still only just beginning,” Dan explains.
Sharemind is pretty hardcore IT-science, and science is not usually perceived as very startup-esque. “Young people often see science as something dusty or slow-paced, whereas you want something dynamic, passionate and groundbreaking. Look, I never wanted to be a scientist. Being a scientist just allows you to do the things you want to do, and literally change the world,” Dan says.
From Tartu to DARPA
Despite the research having gone on for years, the true journey began when Dan joined Cybernetica in 2007 and compiled a team of engineers and researchers. The real breakthrough was still a couple of years away.
“By that time we had been programming and researching this for some years already and we started to understand that we might be uncovering something big,” Dan says.
“Of course we were hoping for a breakthrough, to get a contract for something real, with real life value,“ he continues.
In 2009, Sharemind got its first contracted customer. It was DARPA (The Defense Advanced Research Projects Agency) of the U.S. Department of Defense.
“I think for a lot of people, including our team, it is still a little bit hard to believe it. I still remember the first project meeting in the States you step into a huge room, filled with the greatest brains on the planet, people who created the Internet, people who created the @sign, people who are solely responsible for some of the most relevant high-tech miracles of today; and then our small team,“ Dan reminisces.
Seven years later, Sharemind is going strong. Its pilot projects have returned positive results, its team has been rapidly growing and the use cases expanding daily.
“Any tech startup knows that in the beginning, you try to build the market to the product. We have been very good at creating solutions, now the challenge is to create the problem as well: identifying those who can benefit and applying a profitable business model to it.”
He no longer writes the code for Sharemind, as he is too busy travelling across all continents to introduce the technology and identify new usecases, creating business from science.
“I’m hoping one day you’ll read about it all in a book,“ Dan says with a smirk.