SocGen Builds Serverless-Code Environment, Explores Virtualized Computing

Societe Generale has built a serverless-code environment internally to allow non-developers within the bank to code simple solutions.

The coding environment functions similarly to Amazon Web Services’ (AWS’) Lambda and Google Cloud Functions, which are event-driven serverless computing platforms that run code in response to events. Sumanda Basu, data management technology leader at the Paris-based bank, said the coding environment is being used to democratize technology use and enable non-technologists to develop basic code or analytics functions.

“It helps [employees in the bank] migrate from being an end user, computing [on] Microsoft Excel, [to] move to more IT managed applications—democratizing this technology usage,” he said, during a panel discussion on technology trends and strategic planning at the inaugural WatersTechnology Innovation Exchange.

The expectation is the ability to use the technology should be with every function in the business.

Sumanda Basu, SocGen

Programming and automation skills have become increasingly important in financial services roles over the years. According to a recent report from the global recruiter Robert Walters, a third of all job vacancies at banks are related to technology.

Echoing the research, Basu said banks are scouting talent with technology skills, even for traditional finance or risk roles. “The expectation i is the ability to use the technology should be with every function in the business,” he added. The serverless-code environments are deployed to help fill the skills gap and lower the barrier to entry for non-tech natives to be able to write solutions.

SocGen is also conducting an internal proof-of-concept (PoC) using virtualization. The bank is exploring virtual computing to better manage the firm’s tech operations and avoid migrating or moving around parts of its data stack, including those that are on premise, in the cloud, and in its data lake.

“We collected almost 10 different use cases and then [trimmed] down the use cases to something more manageable,” Basu said. “We worked with vendors on the PoC to find out what kind of technology and what kind of processes we can or cannot use for virtualization.”

The bank is working with several providers on the project, but could not disclose the third-party names. Some of the biggest players in virtualization technologies include VMware, Microsoft, IBM’s RedHat, Google, Amazon and Citrix. For instance, Goldman Sachs is undergoing a multi-year pilot program to modernize its customized virtualization environment using Red Hat.

Like any major project that requires funding, the virtualization PoC had to seek clearance and approval from SocGen’s senior executive committee, Basu said. And convincing those who hold the purse strings to invest in new technologies remains one of the biggest internal struggles at firms today.

Technical Buy-ins

According to Basu, cultural issues within institutions are a leading cause of failed projects. Not only is there difficulty in getting approval for a project from the C-suite, but tech leads must also wean internal teams or clients off of legacy systems and practices.

“We need to resolve these cultural issues; we need to have the clarity of vision and educate [staff and senior executives]. I have seen some of the approaches and initiatives in our company and other companies I have worked with before fail because of this problem,” he said.

Sumanda said everything begins with cultural change. For instance, technology integrations can be significantly problematic because of data inconsistencies and various product names. For example, he said a bank’s front office might follow Bloomberg or Refinitiv’s product structure, whereas the back office might align with Broadridge, SunGard (FIS), or Shadow Financial Systems’ product hierarchy. Tackling problems like these would ultimately require a shift in cultural thinking, not only within a single institution, but also across the capital markets, in creating a common data language.

“How you align all these things like semantics, taxonomy definitions, connecting it syntactically and semantically is a huge challenge.These are just some things we are all working toward together to have a more framework-based approach,” Sumanda said.