Good Help is so Hard to Come by After All
As of the writing of this article, the popular recruiting site Indeed.com returned 1745 search results for the query “Data Scientist Financial Services” (If you’re in data science, hang with us for a second before you scurry over there). Google returns pages
upon pages of results when prompted to find the secret to “recruiting data scientists.” The words “war” and “drought” are used. This is not a drill. With so much talk about the shortage of data scientists, John Ginovsky of Banking
Exchange helps us draw the obvious conclusion. “Data remains a big deal in banking, and its essential corollary, data analytics, promises to transform the business positively in all the ways that matter: security, compliance, customer satisfaction, and,
not incidentally, profits. The problem that’s becoming increasingly serious: While tools for collecting, sifting, and sorting data become faster, cheaper, and better, people with the skills to decide how to make use of the results grow harder and harder to
Great. We agree on the fact that skilled data scientists can positively impact your financial institution’s bottom line, but they’re difficult to find. How do we operate in that reality? Well, what if all of these tools that make collecting and sorting data
so simple could also help fill the talent gap that FS companies are facing? That’s where cloud-based machine learning services come in. While job descriptions for most data scientists take a cue from NASA’s senior leadership profiles, a tool like Amazon Machine
Learning can open up opportunities for junior or internal hires to augment your risk analytics team, provide immediate value, and grow into more advanced roles. For example, with Amazon Machine Learning, a credit risk model can be trained and deployed in about
20 minutes, by someone with little to no experience.
Machine learning might not be just a temporary solution to a talent problem. McKinsey & Company studied more than a dozen banks in Europe that have replaced older statistical-modeling
approaches with machine-learning techniques and saw increases in sales of new products, savings in capital expenditures, increases in cash collections, and declines in churn.