Just because one Goldman Sachs MD quit banking for an elusive residency at Google Brain, don’t assume you need to leave finance to pursue your passion for machine learning. As artificial intelligence (AI) comes of age, there are a growing number of firms putting self-teaching computers to work in financial services. Better still, most of them are hiring.
These are the top companies in the finance world for machine learning right now, along with who they like to hire and the vacancies they have open. If you’re interested in machine learning, you may want to apply soon – universities globally are currently training-up thousands of machine learning specialists, including several savvy former traders who saw this trend coming.
What is it? A ‘revolutionary, award-winning [financial services] search engine that helps you to instantly cut through the noise and uncover critical data points that others miss.’ Alphasense allows users to search in-house content and company disclosures alongside external sources like sell-side research. It already has around 500 clients.
What does the AI do? Alphasense offers “professional search functionality”, which it says “leverages sophisticated natural language processing and search technology that streamlines finding and tracking the most relevant information.” Basically, it employs intelligent search and language algorithms that learn from past mistakes and successes and make the process of unearthing relevant data more efficient.
Who is AlphaSense? The company was founded by Jack Kokko, who was once a former analyst at Morgan Stanley. Kokko spent three years at Morgan Stanley between 1997 and 2000 after graduating from the Helsinki School of Economics. Its EMEA sales staff include a former professional golfer. Several of its staff come from S&P or Factset Research systems, both of which provide competing (albeit non-AI) intelligence products. Alphasense says it looks for “outstanding candidates” across, “computational linguistics, search engine technologies, cloud computing, and securities research.”
Who should be afraid? Financial journalists (who’ll be superannuated by an excellent search engine), buy-side researchers, possibly.
What is it? Cerebellum is a hedge fund management firm which uses an AI system to make investment decisions.
What does the AI do? In the words of Cerebellum itself: “The system is responsible for constantly creating its own new models for how the markets will move, testing those models, refining them, and learning trading strategies that take advantage of these predictive models.”
Who is Cerebellum? The CEO and director is a Stanford PhD with a long history in machine learning (including an algorithm based on physiological data). The COO is a former partner at ‘TrimTabs Asset Management.’ The company also employs portfolio managers and strategists at its HQ in San Francisco. It would like to hear from, “world class professionals who genuinely enjoy crafting, tackling, and solving the challenges associated with a fully automated analysis and trading system.”
Who should be afraid? Quantitative hedge fund managers, maybe all active stock pickers.
What is it? Something that “transforms” Twitter streams and other “public datasets” into “actionable alerts” that investors can act upon.
What does the AI do? Dataminr uses an algorithm to analyze Tweets and other publicly available data so that it can pick up on breaking news before it’s reported elsewhere.
Who is Dataminr? The company was founded in 2009 by three undergraduates at Yale: Ted Bailey, Jeff Kinsey and Sam Hendel. Bailey and Kinsey have only worked for Dataminr, but Hendel describes himself as co-founder for Dataminr whilst holding down a day job as an event driven hedge fund manager at Levin Capital Strategies. Dataminr has around 200 staff and offices in New York City, Seattle and London. It’s currently looking to fill around 30 roles, including one for a recruiter in London (suggesting more hiring is coming soon).
Who should be afraid? Desk-based researchers in banks and hedge funds.
What is it? iSentium “extracts sentiment” from social media information and “transforms it into actionable indicators.” Clients include hedge funds and investment banks.
What does the AI do? iSentium uses in-house algorithms to decipher and process millions of social messages and assign each one a sentiment score. The models are not self-learning however: iSentium explicitly says that there is no machine learning involved in its product. Instead, it relies upon, “artificial intelligence structure-dependent technology,” which, “makes correct predictions on the sentiment of short texts, such as tweets, where natural language constituents are missing, as well as longer texts, which may include non-relevant topical information.”
Who is iSentium? The CEO and chairman previously worked for NEC’s super computer division. COO, Sameer Gupta, was the COO for global electronic trading and Americas high touch and program trading equities business at J.P. Morgan until he joined iSentium in 2014. iSentium is advised by David Hellier, former co-head of the securities division at Goldman Sachs. It has offices in Miami, New York and Montreal and is currently hiring a data engineer and data operations associate.
Who should be afraid: Strategists, researchers.
What is it? Kensho uses “cloud based software” to scan documents on everything from drug approvals to economic reports, monetary policy changes, and political events and to produce answers to “more than 65 million question combinations” on where markets are headed. Eg. ‘Which cement stocks go up the most when a Category 3 hurricane hits Florida?’ Investors include Goldman Sachs and Google. Kensho’s founder, Daniel Nadler, told Goldman Sachs that he came up with the product after being astonished to discover that there was no way of efficiently assessing the impact of geopolitical events on markets beyond looking at past events and manually creating some spreadsheets.
What does the AI do? Kensho uses natural language processing (NLP) systems that are able to learn how to look for new and unexpected relationships between events and asset prices and to recommend searches users might not have considered.
Who is Kensho? 33 year-old Nadler has a Havard PhD and has worked as the director for financial research at Stanford University’s school of digital engineering. Kensho is filled with ex-Google engineers and the chief operating officer is a former managing director and head of front office computing at Credit Suisse. Kensho likes to hire PhDs and is currently looking for a software engineer based in its testing department in Cambridge Massachusetts.
Who should be afraid? Salespeople at Goldman Sachs. Goldman’s salespeople are already reportedly using Kensho to respond to client requests. In future, the likelihood is that clients will simply access Kensho (or ‘Warren’ as its interface is called) themselves. As Nadler himself points out, Kensho can do in minutes what someone earning $350k could do in 40 hours.
What is it? Toronto-based Quandl is an ‘open data service.’ It provides free data directly to people who work for hedge funds, asset managers and banks and make trading decisions. Quandl’s big selling point is its ability to also provide these investment professionals with “alternative”, “alpha-generating” datasets which they can’t typically access through traditional sources, for a fee.
What does the AI do? Quandl uses AI to look for undiscovered data and to evaluate its relevance.
Who is Quandl? Quandl’s co-founder and CEO, Tammer Kamel was once an analyst at Citi (back in 1995). Its other co-founder has a background in rates trading. Quandl is hiring across marketing, engineering and data for its office in Toronto.
Who should be afraid? Bloomberg and Reuters. Quandl offers a huge amount of easily interpreted and manipulable data for free.
What is it? An investment company using artificial intelligence to develop proprietary quantitative trading and investment strategies.
What does the AI do? Like Cerebellum Capital, Sentient uses machine learning to evolve and optimize its trading algorithms.
Who is Sentient? Sentient’s chief investment officer, Jeff Holman, was formerly chief risk officer at Highbridge Capital Management and began his career at Citadel Investment Group. Its CFO is a former convertible trader at Gabelli Asset Management and the head of its platform formerly worked on algorithmic execution strategies at Citi. Sentient has offices in San Francisco and Hong Kong and is currently hiring algorithmic traders, researchers, engineers and salespeople. It’s running an internship for PhD students in summer 2017.
Who should be afraid? Quantitative hedge fund managers (see 2).