– If I have two children, one of whom is a girl, what is the probability that both children are girls?
For most people, this type of question will remain a mystery. However, if the answer is obvious to you, then you might be one of the 20 people who will get a job with American Express in the next 12 months. (AmEx) as a data scientist.
‘Solving this problem is no different from how we create a machine learning model and how we develop a mathematical model to solve real problems,’ says Di Xu, vice president of AI Labs and the AmEx’s artificial intelligence management department.
The company’s activities in the field of recruiting employees for data scientist positions reflect the general increase in industry demand for data experts. According to the Bureau of Labor Statistics, in 2020-2030, the number of jobs in data-related positions will increase by 31.4%. AmEx’s shift from using traditional statistical models to machine learning is synonymous with the growing presence of artificial intelligence (AI) technology in business processes.
Amercian Express began investing in artificial intelligence as early as 2010, says Anjali Dewan, vice president, consumer marketing and personalization business decision making. The company hopes the technology will help process eight billion AmEx transactions.
See also: Humans won’t be able to control the super-intelligent AI. These are the findings of a new study
The decision to implement AI technology turned out to be the right one. Since 2014, 100%. the company’s credit and fraud departments use artificial intelligence.
When interviewing data scientist candidates, recruiters want to know a person’s skills, ability to learn and adapt. These features are essential in efforts to improve the quality and effectiveness of artificial intelligence.
It’s the approach that counts
Recruitment interviews are divided into two types. One of them deals with technical issues, the other examines behavioral skills. As Xu points out, the questions are mostly open-ended, through which company representatives can observe the process of finding a solution presented by a given candidate.
“The final answer, good or bad, is less important to us,” Xu says. – We are more interested in how the potential employee intends to solve the problem.
Although some questions cannot be answered unequivocally, these problems can be solved in different ways. – Sometimes we ask additional questions along the way to see how the person approaches it and what their response to the question will be – he adds.
At the very beginning, candidates are often asked a slightly simpler question, such as:
– If I have two children, one of whom is a girl, what is the probability that both children are girls?
Some people will look for an answer to this question using a mathematical theory that can help find a formula; others will use computer code to find the answer. However, Xu points out that the process of finding answers is more complex.
See also: China is ahead of the United States in key technologies of the future
“It is possible to answer this question in a very simple and concise way,” says Xu. – If a person is able to answer yes, then it is no different from creating a mathematical model aimed at solving a given problem.
In the next step, the candidate receives a similar question, extended by one dimension. – I have two children, one of whom was born on Tuesday. What is the probability that both children are girls?
Providing the candidate with a well-phrased answer to this more complex question helps to understand how the person approaches the issues presented to them.
Other Questions Applicants May Receive work in AmEx are more hypothetical, for example you have some kind of data product to design. We need to know the entry prices in each restaurant. How will we use your transaction data for this purpose?
Transaction information is one of the most important data for AmEx because it reveals many important issues and enables the company to develop innovative customer solutions, Xu says.
The vice president of AI Labs admits that he himself doesn’t know the perfect answers to hypothetical questions. Instead, he prefers to hear the candidates’ ideas.
– Perhaps they will be able to give us ideas, thanks to which we will get the perfect solution on a given subject – he underlines.
The key desire to develop
In the behavioral part of the recruiting process, recruiters test candidates for leadership qualities, Dewan explains. A potential employee is expected to work effectively in different teams to better understand the problem.
Dewan and Xu agree that the most important thing is a person’s ability to learn and willingness to grow.
– Competence is only one aspect. The desire to continue learning is another equally important issue. Xu pointed out. – This is a key factor for future success.
Crowd. Adam Hugh