Are we looking for an algorithm that “If we all die, it would keep trading”? Should we be worried that electronic trading is mushrooming like airplane traffic, while we are not paying that much attention?
Today, I’ll look for AI pigments of incremental changes in algorithmic trading, first on Wall Street and then outside, in the Fintech startup world. I am not including the HFT space because it is a particular space driven by speed and merits a separate post because of its politically sensitive angle (Michael Lewis’s babe).
Renaissance Tech and Two Sigma, are probably the most recognizable names in old fashioned quant trading space. Bridgewater, publicly announced a year ago that they will start a dedicated AI team reporting to David Ferrucci, who joined Bridgewater 3yrs ago after leading the IBM group that developed Watson. The FT reports that Mr Luo chief quant at Deutsche Bank, leads a team that has developed an AI algorithm that searches the financial system for investment opportunities, “scraping unfathomably large data sets to unearth profitable patterns to proffer clients”.
Early stage startups focused on using advanced quant techniques for trading are mainly focused on machine learning techniques to calibrate models that manage money typically on a discretionary basis.
Walnut Algorithms out of Paris, is applying latest advances in data science and machine learning research to the financial markets. They combine advanced machine learning techniques with financial expertise to generate medium to high frequency pure alpha strategies. Euklid, out of Italy, employs a sophisticated AI engine capable of making independent trading decisions. The AI is based on complex algorithms which are monitoring technical indicators and analyzing market trends. Both are StartupBootcamp Alumni. Euklid has a much broader vision of becoming the first AI bank, offering digital crypto currencies and employing smart contracts.
ArtQuant, out of Moscow, has created a product PortfolioandMe, positioning AI as a service for portfolio advice, monitoring and construction. If the customer chooses the in-house portfolio construction, then the AI algorithms of ArtQuant are used with a fundamental analysis approach.
Similar to a hedge fund approach, Aidyia, from Hong Kong, is taking AI in finance to the next level, acting as an asset manager with a long short equity fund. Aidyia plans to launch more fund products using quant techniques from genomics and robotics.
Another company taking general business intelligence algorithms and applying them to financial trading, is Sentient technologies, a San Francisco company who got more than $100mil funding at the end of 2014. Sentient has been trading with AI systems over the past year (no fund product available yet) operating in the discretionary management space. In last week’s article in BBC news, Would you let a robot invest your hard-earned cash? Sentient AI was mentioned as one of the two companies focused on training algorithms to learn from past mistakes and refine their rules, without the need for much human intervention. New-York based Rebellion Research is the other one, which is offering machine learning asset management. They offer a basic Global Equity strategy, which has been managing money since 2007 via a Hedge Fund structure and as well as in managed accounts. They also offer the possibility of creating a Market neutral strategy; an absolute return strategy and a US bond strategy. These sophisticated strategies are available through managed brokerage accounts at Interactive Brokers.
Another brokerage firm getting involved in AI more directly is Silicon Markets, a Fin-Tech, that uses MT4 platform and a customized version of it for its brokerage business, Tradable. Silicon markets is launching a Machine Learning Trading Optimizer which will be integrated with Tradable.
Fintechs threating directly professional asset managers, are those also that empower the DIY retail traders and create alpha directly for end users (investors).
Alpha Modus is delivering alpha directly to end-users through a marketplace of investment technologies they call ‘mods.’ These mods are priced according to how much alpha they generate, so investors can buy alpha on this platform without paying the conventional high fees of alpha-generating asset managers. One of the key technologies available in the Alpha Modus marketplace is the Early Look Imbalance Meter, which aims to predict the direction of the NYSE into the closing auction. Recently, Early Look mod has been combined with the IBM Bluemix platform, where they have incorporated Watson’s Insights and Twitter sentiment analysis. A live case study of cloud based collaboration between fintechs and incumbents.
Quantopian of course, belongs in this classification too, since it offers a platform for quant algorithms development, back testing, and also, is creating the first crowd-sourced hedge fund by backing selected investment algorithms.
Capitalico is a trading platform that makes quantitative analysis and algorithmic trading possible for everyone on a smartphone. With no programming skills, the platform allows you to build, test and trade an algorithm out of an idea with a few clicks and based on visuals of historical charts. A powerful tool for DIY traders, out of California. Alpaca is the company behind this platform, been developing Deep-Learning trading algorithms.
Celebrating AI in digital asset &wealth management, continues with ten picks focused on algorithmic trading of all sorts (long-short equity, crowd-sourced hedge fund, mods, etc), to sophisticated enablers of DIY individuals. AI in digital wealth management is also morphing customer preferences and profiles. These incremental services are changing trading and investing by blurring the lines between these activities.
Daily Fintech Advisers provides strategic consulting to organizations with business and investment interests in Fintech. Efi Pylarinou is a Digital Wealth Management thought leader.
[…] Wednesday continues the focus on the impact of AI on Digital Wealth Management., specifically “… […]
Bridgewater is one of the oldest and probably the biggest quant player in the asset management industry, they simply never marketed their business as a “quant”. Most of the quant players have been accomodating machine learning and big data techniques for a long long time. What these new fintech startups are doing is ridding the “machine learning” and “artificial inteligence” marketing hype. I started in the industry working with unstructured data in 1999 and by 2007 when I left, I was running a fund based on genetic programming. Algorithms are just a tiny part of the equation. I’ve seen many startups like these come and go, albeit with the flavour du jour – systematic, quantitative, etc – and I wish them luck, they’re going to need it to make enough money to survive.