AI Trends Weekly Brief: Boston Fintech Startups

102

Challenging Business Models With Quantitative Innovations

Boston has a robust financial industry, being home to notable firms such as Fidelity Investments and State Street Corp. Also known as a hub for technology and innovation, Boston is now building a reputation for its impressive range of startups in fintech, companies that use technology to support and enable banking and financial services in new ways.

Often collaborating with the legacy financial sectors, the entrepreneurs are creating new ways to invest, pay, save and more. They play in areas including personal finance, cryptocurrency, investment management, trading and lending. Many of them are leveraging artificial intelligence within their products and services. A sample of Boston fintech startups illustrate the trend.

Kensho Seeks to Lead Financial Machine Learning

Kensho uses a huge historic database and machine learning to analyze how a specific event–from a natural disaster to an election result–might affect markets, presenting results in an easy-to-digest knowledge graph. The firm’s approach combines natural language search queries and secure cloud computing to create analytics tools that a wide range of investment professionals can use.

Kensho was founded in 2013 and has raised $67 million from investors, including Google Ventures, Goldman Sachs and In-Q-Tel, the venture capital arm of the CIA. Founded out of MIT and Harvard, Kensho’s team comes from veteran positions at Google, Apple, Facebook and academia.

The team includes one of the seven members of the original Apple iPhone engineering team, PhD physicists and quantum computing PhDs. Its products are used by Goldman Sachs, JP Morgan Chase and Bank of America Merrill Lynch.

Cofounder and CEO Daniel Nadler, 33, wrote his thesis for a Harvard economics Ph.D. on sovereign credit risk.

Kensho’s machine learning systems crawl through reams of data and market-moving information, searching for correlations between world events and their impact on asset prices. The Kensho Global Event Database, one of the world’s largest repositories of data and information, powers its machine learning analysis. Kensho’s Knowledge Graph acts as a real-time living graph model of world events. The platform’s information is packaged with intuitive search tools and data visualization capability.

“We have very long-run ambitions. We want to be the leader in machine learning for the financial industry for the next twenty-five years,” CEO Nadler told Forbes in an interview.

“The coming era will be looked back upon as the ‘AI era,’ when AI became the defining competitive advantage for corporations, government agencies, and investment professionals alike,” Nadler stated in a press release. “The combination of S&P Global Market Intelligence’s high quality data and commercial reach with Kensho’s advanced machine learning capabilities will serve to position both companies for future growth and success,” Nadler stated.

Quantopian Crowd-Sourced Quants Get Paid

Quantopian is a Boston-based company that aims to create a crowd-sourced hedge fund by letting freelance quantitative analysts develop, test, and use trading algorithms to buy and sell securities.

The firm was founded in 2011 by John Fawcett, CEO and Jean Bredeche, CTO.

DIY quants create investing algorithms using the site’s data and research. The best performers are licensed for a 10% cut of profits. Hedge fund billionaire Steven A. Cohen has said he will invest up to $250 million in top strategies. Funding of $25.8 million has also come from Bessemer Venture Partners, Spark Capital, Khosla Ventures, Cohen’s Point72 Asset Management and others.

Quantopian in April allocated tens of millions of dollars to authors of some of the best of its crowd-sourced algorithms. The allocations ranged from $100,000 to $3 million per algorithms. These are the first allocations made with external money being managed by Quantopian.

The freelance quants receiving money were from eight countries including Australia, Canada, China, Colombia, India, Spain, and the United States. They came from many walks of life, including data science, finance, engineering, software development, and academia.

Each author is paid 10 percent of the algorithm’s net profits and retains ownership of the intellectual property.

“Talent is everywhere in the world, not just at hedge funds located around New York City,” stated co founder John “Fawce” Fawcett, CEO, in a press release. “Quantopian gives talented people anywhere the chance to learn how to create new investing strategies and apply what they’ve learned on our platform. We are just beginning to give people a chance to put their knowledge to work.”

Quantopian selected the algorithms from the Quantopian community after a rigorous process with both automated and human components. Quantopian intends to ramp up both the size and the number of allocations on a monthly basis.

“We are looking forward to increasing steadily the value of our allocations in the coming months — from tens and then to hundreds of millions of dollars,” said Jonathan Larkin, Quantopian’s Chief Investment Officer.

Quantopian’s unique platform allows authors to research and test their ideas and use a fast-growing catalog of high-quality datasets. More than 120,000 members from 180 countries have used the platform and run more than 6 million simulations.

“Our research and educational platform for algorithm authors is now matched by a state-of-the-art trading system to provide high quality trade execution services for investors,” Fawcett said. Quantopian has assembled an experienced trading team over the last year as part of its investment management practice.

How it Works

Quants, also known as authors, are provided with everything they need to create their algorithm. Quantopian operates in an open-ended Python environment, offering free access to stock price history, consolidated databases, corporate fundamental data sets, and several APIs, among other resources, all for the purpose of empowering the quants.

The startup also provides integrated development and execution environments that allow for the quants to test their algorithms for free, then move it into a real-world setting to start informing trades.

After the algorithm has been created and deployed, Quantopian evaluates each individual’s algorithm and, if the author agrees to license his or her algorithm to Quantopian, allocates funds accordingly to the best based on the algorithm’s capabilities and features. Quantopian manages these algorithms for the benefit of investors in its vehicles, essentially acting as the intermediary between the quant and investor.

A piece in Inc. Magazine by Alex Moazed, founder and CEO of Applico, stated, “Quantopian demonstrates a strong potential to alter the future of hedge fund management by carving out an original role in an increasingly tech-driven financial industry. The industry is undergoing radical changes as it transitions from a traditional discretionary model toward highly automated, systematic, quantitative investments.”

“There is a bright path ahead for Quantopian as its business model is completely unique, not found anywhere else. There are quantitative hedge funds in place, but none of them have a comparable platform community powering the decision-making. Quantopian plans to scale over the rest of the year.”

Reuter’s QA Point Based on Elsen’s Platform

A Boston-based financial tech startup that has raised less than $1 million in investment funding got a big boost recently when Thomson Reuters unveiled a new data tool based on its technology.

Thomson Reuters Corp.’s QA Point, a software tool that lets investment analysts test their investing models against historical financial data, is built on local startup Elsen Inc.’s flagship product, nPlatform, which Elsen also recently launched.

Three Northeastern University graduates founded Elsen in 2014 and the fintech company participated in the Boston-based nonprofit accelerator FinTech Sandbox in 2015. Elsen has raised $750,000 from Hyperplane Venture Capital, serial entrepreneur Bret Siarkowski, and an Accomplice-affiliated syndicate of angel investors.

Zac Sheffer, co-founder and CEO of Elsen, said he saw the need for a more technologically advanced way for investment firms to test their strategies against sophisticated financial data during a stint at Credit Suisse.

“From a financial standpoint they really know their stuff, but from a technology standpoint they’re decades behind where they need to be. At that’s really the standpoint for the industry as a whole,” Sheffer said in an interview with the Boston Business Journal. “They’re still trying to manage a $100 million portfolio in Excel.”

QA Point is designed to assist in creating increased productivity, better results and accelerated internal collaboration among investment professionals. The tool uses a point-and-click interface and access to Thomson Reuters content.

Many portfolio managers seek to use quantitative models to optimize the risk/return profile of their investment strategies, but do not have the highly technical “quants” needed to do the analysis work.

By offering the analysis in the cloud, QA Point seeks to add value immediately, and make it easy to share models and strategies within an organization’s research team. No time is spent installing software on premises; QA Point is accessed via a web browser.

“Now more than ever, asset managers need cost-effective, transparent solutions that drive collaboration throughout the investment management process,” stated Pradeep Menon, Managing Director, Global Head of Advisory and Investment Management, Thomson Reuters. “QA point gives asset managers tools that were typically only possible with highly skilled developers and quantitative analysts, enabling them to take advantage of investment techniques that help deliver results for their clients.”

With QA Point, users can access a wide range of data and content including Thomson Reuters I/B/E/S and Worldscope Fundamentals, as well as third-party content. All the content and data is integrated into a single, global standardized database with a comprehensive symbology mapping, facilitating easier data management.

Users also have access to StarMine Quantitative Analytics and stock section models, with transparency into the underlying inputs so users can more confidently employ complex stock selection factors into their models.

Once models are created, users can conduct backtesting – often one of the most time-consuming steps in the investment management process – in minutes, not days. QA Point offers a wide variety of statistical measures, strategy analysis tools, built-in factor testing tools, and data visualization which could significantly speed up the time it takes to backtest new models.

QA Point was created on the Elsen Platform, designed to give financial institutions a foundation to build web-based applications that allow users to more easily harness, understand and help make quick decisions with vast quantities of data without a team of expert programmers.

“Thomson Reuters is one of the most respected sources for financial news and data in the world, and has never stopped innovating new ways to help its clients drive more value from its data,” stated Elsen CEO Sheffer in a press release. “By choosing to partner and build QA Point on the Elsen nPlatform, Thomson Reuters has once again shown its commitment to delivering solutions on the cutting edge of today’s technology. We’re looking forward to helping Thomson Reuters deliver continued value to its clients.”