You can tell a lot about a business from a visit to their office. A few minutes’ walk from London’s hectic Oxford Street, hidden inside a converted Victorian terraced townhouse turned office building, ASI Data Science deliver one of the most-promising AI offerings to a host of big brands, including the likes of Easyjet, BBC and Tesco.
The building exudes cleverness; blackboards smeared in formulae, a ping-pong table that is actually used, an atmosphere that’s more akin to the friendly offices of a university’s science department than a corporate headquarters.
I was there to learn about the natural career path of a data scientist – something that ASI’s Head of Research, Ilya Feige, knows a lot about.
His first words to me: “I am not a data scientist.” A smile appears on his face, “I work mainly in machine learning and artificial intelligence on the research side, so I wouldn’t consider myself a data scientist as such.”
Like many of his colleagues, the Harvard doctoral graduate comes from a background in physics, and has been enticed over to the business world with high hopes of making a positive change.
“Here there’s opportunity to have a big impact on improving the lives of others.” He blushes as soon as he hears his statement out loud.
“I know, it sounds cheesy,” he admits. “But it’s true. Data science and machine learning automate the mundane, everyday tasks people don’t want to do, but this technology also has the potential to open up access to information and services that are currently out of reach for a lot of people.”
“Take access to quality and affordable healthcare as an example – this is one of the biggest problems facing developed economies. With AI technology, individuals could one day own their healthcare data, and algorithms could diagnose people’s health ailments if they are struggling to get advice from a doctor.”
He’s not alone in his thinking. In the last couple of years, data scientists have made substantial headway transferring medical facts and figures into complex algorithms; the day that AI technology becomes common practice in diagnosing medical ailments is right around the corner.
“Good language and image processing can turn unstructured healthcare data into structured data, which can in some cases be analysed by algorithms more accurately than by doctors. If this came into effect, I think healthcare could be ten times cheaper than it currently is, and better quality too.
“The same principle applies to legal services. When you need to hire a lawyer, most of the time it’s because they know a set of rules that you don’t - those rules should be stored in a machine that’s accessible to everyone.”
THE GROWING DEMAND FOR DATA SCIENTISTS.
Ilya and Tim discussing what it takes to have a career as a data scientist.
As opportunities rise for organisations to improve their offering through data science, the demand for data scientists is growing by the hour. According to Indeed.com’s EMEA Economist, Mariano Mamertino, the number of AI jobs in Britain has soared by 485% since 2014.
“Data science as an industry is extremely new – there’s nobody that’s been doing this for ten years,” said Ilya. “Data science as it is today only came into being over the past three or four years.”
After graduating from Harvard, the Canadian research scientist spent a couple of years at McKinsey & Company in San Francisco before making the move over to London.
At ASI, Ilya splits his time between being the technical lead of the data science consulting team and his own research, which ranges from modelling the semantic meaning in natural language processing through to training undirected neural network models.
“We do very little of the stuff we don’t like doing,” he admits.
MINDSET OVER MATTER.
Originally a fellowship programme, ASI Data Science has since branched out into a talent company of their own, providing organisations with in-depth insights and recent graduates with front-line industry experience.
I ask him what qualifications you need to get into a career as a data scientist.
“Ironically, science degrees are far more applicable to the type of data science we work on than a degree in data science or computer science,” he replies.
“As the industry is so new, and the specifics of a career in data science are still very hazy, when interviewing fellowship students I always make a point of asking: “What makes a good data scientist?” This is always a good indicator of what they care about and are interested in doing.”
At a time where AI experts and PHD students are being routinely courted by organisations on the same scale as Google, Facebook and Amazon, hundreds of applicants are still applying for ASI’s prestigious fellowship programme every year with the hopes of earning one of the 50 places on offer.
“To be an effective data scientist, you need the necessary technical capabilities: strong maths and statistics, a good grasp of the scientific method, decent coding skills (particularly Python, or SQL) - and some knowledge of machine learning techniques is always a plus.
“But aside from the technical, it’s equally important for a data scientist to have outstanding soft skills. A data scientist really adds value to an organisation by communicating complex ideas in an understandable, jargon-free way and to a varied audience. And data science is almost always collaborative – so you also have to be a team player.”
He continues: “And, perhaps most importantly, when you’re given a problem, you have to be able to own it. Rather than simply doing what you’ve been asked, you need to explore a range of different options and uncover all value, and test your model’s robustness. Candidates that can think with their business hat on – like an entrepreneur – will have the most impact.”
Data science has fast-become one of the most innovative, exciting and ground-breaking sectors on the jobs market, and as AI technology and machine learning become more commonplace in a business setting, so too will the demand for collaborative and agile-thinking data scientists.