How to get ahead in quantum machine learning AND attract Goldman Sachs

eFC logo

40 years ago a personal computer cost around $500k in today's money and was accessible only to large corporations. Today, as the cliché goes, that kind of processing power is available to most people in an affordable mobile phone. Quantum computing, however, is a different matter.  Quantum computers are stuck in the 1950s: there aren't many of them, they cost tens of millions of dollars, and they take up entire rooms.

One of today’s very rare and very costly quantum machines is being developed by D-Wave Systems Inc., a company whose CEO happens to be Vern Brownell, a former CTO of Goldman Sachs. Goldman is one of several lead investors in D-Wave, which it’s described as having a “head start” in the field. While most quantum computing rivals are still in their infancy, D-Wave has already been using its system for machine learning. Competitors are eyeing the same plot: 1QB Information Technology Systems Inc (1QBit), a Vancouver-based quantum computing, counts derivatives exchange CME Group among its investors. An RBS banker who led 1QBit's 2015 finance round told Bloomberg quantum computing is perfect for the data-rich time-sensitive world of financial markets.

Interestingly, therefore, an opportunity has arisen to write machine learning algorithms for quantum computers and then implement them using D-Wave 2000Q, the company’s first commercially available quantum computer. Training on the system will be made available too.

The quantum machine learning program is being run by the Creative Destruction Lab (CDL), a seed funding program for science-based companies based in Toronto.  Last month, it invited applications for 40 places on an initiative intended to develop and sponsor a wave of quantum machine learning start-ups. The next (and last) round of applications closes on Monday July 24th.

Daniel Mulet, associate director of machine learning at CDL says they’ve already received 42 applications, around 10% of which are biased towards financial services. “Some are very early stage and have been submitted by students, but others are companies that have already been launched,” says Mulet. “- There’s one that’s working with a hedge fund looking for patterns with trading data.”

Traditional computers use binary code to solve problems: a bit can be a 1 or a 0. Quantum computers use qubits: a bit can be a 1 or a 0 or a 1 AND a 0  As Bloomberg points out, therefore, if you have two qubits you can have four potential states: 00, 01, 10, and 11. Moreover, the number of states a quantum computer can take into consideration is 2 raised to the power of the number of qubits: if you had a 50-qubit universal quantum computer, you could explore 1.125 quadrillion states simultaneously.

“Quantum computers are able to process much larger quantities of data much faster,” says Mulet. “It’s our belief that these new quantum hardware platforms built by D-Wave or IQB will be used for various machine learning applications in the next few years. When that happens, we want to be ready to leverage that. - One day all Bloomberg terminals will be run on quantum computers.”

It’s not hard to see why Goldman is interested.

If you're interested too and want to apply, you have 39 days to polish your application.  As a further lure to candidates, those selected will be mentored by the likes of William Tunstall-Pedoe, a Cambridge AI entrepreneur, and Barney Pell, chief strategy officer at San Francisco-based Loco-Mobi (which is applying AI to parking your car). Those graduating from the program, which begins in September, will receive $80k in funding in return for 8% of the equity in their company.

Mulet says ideal applicants will have a Masters or PhD in a quantitative subject, and be proficient in programming in Python and the use of Tensor Flow, Google's open source library for machine learning. 


Photo credit: Quantum foam by Alex Sukontsev is licensed under CC BY 2.0.

Related articles

Popular job sectors


Search jobs

Search articles