Kirat Singh and the business of sharing Wall Street’s machinery

In 2013, Kirat Singh and Mark Higgins met at Connolly’s Pub in midtown Manhattan to talk about financial technology. Between them, they had worked on the machinery behind trading at Goldman Sachs, JPMorgan and Bank of America. Their proposed next step was unusually sociable for Wall Street infrastructure: let other firms use the same kind of tools.

A pub is a forgiving place to propose a company. An investment bank is a less forgiving place to prove its software. Singh had spent years in the latter. He knew the gap between a clever calculation and a system that colleagues could use, extend and trust. Beacon, the company he and Higgins founded in 2014, would make a business out of that gap.

By April 30, 2025, Beacon belonged to Clearwater Analytics. The acquisition had been announced at approximately $560 million. Singh joined Clearwater as President, Risk & Performance; its current leadership roster calls him President, Risk and Alternative Assets. The titles changed. The underlying preoccupation remained remarkably consistent: give people a clearer view of what they own, what it is worth and what could go wrong.

A mathematics student finds the trading floor

Singh’s route to that Manhattan pub began in India. A Harvard scholarship brought him to the United States in 1992, aged 17, to study mathematics. He graduated in 1997 and joined Goldman Sachs, working on the J. Aron commodities trading floor in New York. Software had interested him from childhood. Here was a place where writing it had immediate commercial consequences.

The setting matters. On a trading floor, an analytical problem arrives with a customer, a transaction and a deadline attached. A model has to meet the untidiness of actual business. Singh later described his early Goldman environment as entrepreneurial. It was an apprenticeship in making ideas usable, surrounded by people who wanted answers quickly and could tell when they had received an unhelpful one.

At Goldman he worked on SecDB, the firm’s shared platform for pricing and risk. His professional biography describes eight years building the system. A common framework allowed work done in one part of the bank to become useful elsewhere. The interesting invention was partly organizational: programmers could contribute to shared machinery while pursuing the particular needs of their desks.

Three banks, one accumulating education

In 2006 Singh and Higgins took their experience to JPMorgan and built Athena. Python and C++ were central to the platform. In 2010 Singh moved to Bank of America Merrill Lynch, where he launched and delivered Quartz. Beacon’s later conference biography described Quartz as serving more than 5,000 developers across the bank.

Those names can make the career look like a procession of software projects. They also describe a repeated test: could an approach learned at one institution work inside another? Each bank had its own businesses, colleagues and existing systems. Moving between them gave Singh experience beyond maintaining a successful inheritance. He had to help establish the conditions in which other people could build.

There is an unglamorous discipline beneath that job. Someone has to organize data, manage computation and arrange how changes reach users. Each individual team may prefer to spend its time on a new trade or a better model. A shared platform takes responsibility for the repetitive work, while supplying the controls that let many teams contribute without turning the whole operation into an elaborate group project.

The business hiding in the plumbing

By the time Singh and Higgins were discussing their own venture, they had seen the limitations of outside software as well as the possibilities of bank infrastructure. Firms could buy individual products and still end up responsible for a complicated patchwork. A vendor’s application might solve today’s problem while making tomorrow’s modification surprisingly difficult.

Beacon’s proposition put a developer environment beside trading and risk applications. Customers would receive useful capabilities immediately, with transparent, licensed source code they could inspect and extend. The distinction is practical. A customer with an unusual instrument or a proprietary model needs room to change the software, rather than simply wait for a vendor’s next product meeting.

That transparency also needs boundaries. Access to a platform’s code does not imply access to somebody else’s confidential positions or research. Beacon’s design separated and permissioned client data and source code. The commercial ambition was to spread useful infrastructure while preserving the private work that made each customer’s business its own.

“Our vision is to create an ecosystem where Beacon provides the backbone, and clients and vendors can share solutions.”

Kirat Singh, 2020

Singh compared the ambition to a common delivery platform for applications. The attraction was the possibility of reuse: a developer should be able to concentrate on the part that mattered to the business. It is an appealingly economical idea. A trading firm may have strong opinions about a valuation; it has fewer reasons to feel proprietary affection for a job scheduler.

A platform earns its place through other people

Beacon received Risk’s fintech start-up of the year award in 2017. Its early users illustrated why flexibility mattered. Axpo US used the platform for risk management and pricing, including structured transactions. Asymmetric Return Capital used it for complex derivatives analytics. These were different businesses asking the same infrastructure to accommodate their particular calculations.

Singh was also taking the argument beyond New York. On June 21, 2017, he spoke at the first Tokyo FinTech Meetup, held at NYU’s School of Professional Studies. More than 50 people attended. Beacon had expanded into Tokyo that year. For a company selling shared tools, meeting a new community of potential users was part of the work.

The 2023 Beacon developer conference makes that community visible in another way. Singh was listed for the opening keynote. The programme then moved to customer and partner presentations, including a PIMCO session on moving beyond spreadsheet constraints and a KWA Analytics session on delivering applications. A platform’s story eventually has to include the people who can do something useful with it.

What the customer gets to keep

Blackstone’s use of Beacon with Snowflake offers a concrete example. Snowflake supplied a centralized data environment. Beacon supplied a common place to develop and maintain models, with cloud computing to run them. Market data could feed financial objects used for interest-rate curves, pricing and risk analysis, with results available across different applications.

The business value sits in the connections. An analyst working in one interface and a colleague working in another can draw on the same underlying model. When a calculation produces a surprising result, the relationship between data and model can be examined. That is a more useful kind of transparency than a dashboard that merely looks reassuring.

It also helps explain Singh’s persistence. After years spent helping banks develop their own platforms, Beacon allowed customers to retain their particular analytics while relying on shared foundations. The product carried a lesson from his banking career into a commercial setting: scarce development time should go toward the questions a business actually needs to answer.

The engineer with a dry aside

Singh’s public comments suggest a developer’s attention to operational details, accompanied by a willingness to puncture technology enthusiasm. During the 2024 CrowdStrike disruption, he recalled administrators carrying recovery CDs around Goldman’s fixed-income floor after a computer worm in the early 2000s. He compared that memory with the manual recovery steps still being required decades later.

The promised arrival of AI received a pointed aside. Meanwhile, he said he had checked Beacon’s infrastructure that morning and found it functioning. It was a founder’s account of his own system, but the priorities were plain enough. New features attract attention. Reliability is what customers notice when everything else stops.

His subsequent writing on secure platforms discussed zero-trust infrastructure and staggered deployments. The subjects fit the career. Once other people depend on your software, the ability to update it carefully becomes part of the product. The interesting day at work is not always the day the customer hopes you are having.

After the deal, the same unfinished work

Clearwater’s acquisition brought Beacon’s modelling and developer infrastructure into a larger investment-management business. At announcement, Beacon had approximately $44 million in annual recurring revenue at the end of 2024. Clearwater’s stated plan connected risk modelling with trading, accounting and reporting, alongside capabilities from Enfusion and Bistro. Completing that integration was an ambition, rather than an automatic consequence of signing the deal.

By February 2026, Singh was writing about the increasingly complicated terms of energy instruments. His argument emphasized models that analysts could inspect and modify, plus valuations fast enough to be useful during the trading day. The setting had widened from bank desks to energy portfolios. The demand for understandable calculations had followed him.

In describing his own values, Singh has emphasized consideration for clients and colleagues, responsibility and respect for their time. Those preferences connect the person to the product without requiring a grand theory of leadership. His work has repeatedly asked how people can share complicated machinery and still get on with their jobs. The conversation that began over beers has kept returning to the working day.