Chevron icon It indicates an expandable section or menu, or sometimes previous / next navigation options. HOMEPAGE

The code for Goldman Sachs' internal data platform is now open for anyone to use. The bank's data chief tells us why the firm decided to share something seven years in the making.

David Solomon goldman sachs
Goldman Sachs CEO David Solomon speaks at the 2019 Milken Institute Global Conference Michael Kovac/Getty Images

  • After a six-month pilot period where banks like Morgan Stanley and Deutsche Bank put it to the test, Goldman Sachs' internal data platform Legend is being released to whomever wants to use it on coding-collaboration site Github Monday.
  • The goal is to help standardize data languages across the industry, freeing up time for software and data engineers to focus on bigger projects.
  • "We think of this platform as something that hasn't been available to the world before," said Goldman's co-chief data officer Neema Raphael.
  • Visit Business Insider's homepage for more stories.

Goldman Sachs — and its other major banking competitors — are not often known for their ability to collaborate. 

While the banks may team up on regulatory issues or other industry challenges, Goldman and its peers are often battling each other for an edge across different businesses. 

But data standardization and security — two decidedly non-sexy topics in the world of finance — has become so paramount that Goldman is releasing the code for its internal data platform, known as Legend, after seven years in the making, so others in the industry can use it and build off of it. 

The reason being, according to Neema Raphael, the bank's co-chief data officer, is the "massive edge" from this platform does not come from the baseline code itself — "it's what you build on top of it."   

Data often comes in all shapes and sizes and takes a lot of time and effort to make it into something usable, but with financial services migrating closer to speaking the same data language, putting that information to work will become that much easier.

"We think of this platform as something that hasn't been available to the world before," Raphael said in an interview with Business Insider. He initially announced the firm would put the platform on the cloud of non-profit Fintech Open Source Foundation (FINOS) last year at the foundation's conference in New York.

See more: Pricey data, slashed fees, and poor returns are hurting hedge funds' margins —and some are getting in the business of helping their rivals

The platform, originally named Alloy, and the language, known internally at the bank as PURE, are now collectively called Legend, according to a release on the release. The code will be on coding-collaboration site Github, and banks like Deutsche Bank, Morgan Stanley, Wells Fargo, and RBC have already been using the platform in a six-month pilot period. 

"This is really a culmination of the work we have spearheaded over the last few years," said Gabriele Columbro, the founder of FINOS.

As banks have become more technology-focused, the ethos of Silicon Valley has seeped into Wall Street. Last year at the conference, several executives from firms like BNY Mellon and Morgan Stanley stated that the competition isn't other banks, it's the collective challenges they face — and collaboration is the best way to tackle them.

"You don't need to fight for talent, you need to collaborate for talent," said Donald Raab, a managing director at BNY Mellon, at the conference last year.

Pierre De Belen, a Goldman engineer who was the head architect of the platform, said Goldman hopes that the "simplification" the platform can bring to many data processes will help engineers turn to "focus on real problems." Raphael mentioned that this is a part of the culture change within the firm — and finance more broadly — to make engineers feel valued "like first-class citizens." 

See more: Data scientists and engineers are leaving Amazon and Facebook for hedge funds. Here are the firms that are winning the battle for top tech talent.

The platform, while built for financial services firm, is not limited to the sector. Raphael sees applications for any firm with data silos and demanding regulators.

While companies with limited data engineering or coding chops might be intimidated by taking on such a challenge, De Belen said one of the critical things about Legend is that it doesn't require an engineering degree to use. It's how Goldman was able to access data "fast, easy, and safely" across teams without having to bring in data specialists, Raphael said.

"This is not only something for technologists, but also it's for the securities division and the operations division of banks," said Rob Underwood, chief development officer for FINOS. 

The platform is building off of the International Swaps and Derivatives Association's Common Domain Model for sharing data, and Raphael hopes it can produce "more deliberate collaboration" among financial firms going forward. 

And Goldman plans to keep sharing, De Belen said.

"There's more phases to come to this," he said. 

Wall Street Goldman Sachs

Jump to

  1. Main content
  2. Search
  3. Account