Revant Nayar, Founder
“A Bloomberg report mentioned that 80 percent of AI-based hedge funds were shut down between 2013 and 2018 because the models could not accurately predict regime shifts in the market,” says Revant Nayar, founder of FMI Technologies. “Given the massive complexity in predicting stationarity and regime shifts in the capital markets, AI or other existing technologies fall short of meeting the market expectations.” AI-based models are time-consuming as they demand a huge amount of training data for financial time series analysis. Moreover, traditional AI systems act as black boxes, so faulty AI models and neural networks remain undiscovered. New York-based company FMI Technologies is addressing the limitations of AI through field machine intelligence (FMI) technology. The company offers FMI-based custom solutions for large banks and hedge funds to make accurate predictions and remain profitable.
“FMI is a robust and transparent alternative to AI for financial time series forecasting and classification, anomaly detection, and correlation analysis,” explains Nayar. Developed by senior professors of the physics department at Princeton University, FMI has emerged as one of the most powerful tools to analyze contemporary financial data. It has been mathematically proved to be a unique algorithmic framework that neither mistakes noise for signal nor vice versa. What’s more, FMI-based AI systems no longer remain a black box as all the variables remain under control of the users.
Backed by such technology, FMI Technologies offers ‘algorithm as a service.’ The service is mainly aimed at medium and high-frequency hedge funds. The unique service allows clients to capture the non-linear elements in the market by predicting non-linear market dynamics.
More importantly, FMI Technologies offers multiple versions of its AI model with a clear financial interpretation, so clients can steer clear of modifying the models based on their unique risk level or other attributes. These capabilities drive hedge funds to invest in AI-based solutions with confidence. In addition to FMI models, the company also provides custom-built APIs depending on factors such as frequency and signals that the individual clients deal with.
FMI is a robust and transparent alternative to AI for financial time series forecasting and classification, anomaly detection, and correlation analysis.
At the outset of client engagement, team FMI Technologies examines which trading markets the client is interested in and what frequencies and risk levels they will be trading at. Subsequently, the company provides a custom solution in accordance with the client’s risk appetite. It is because of this personalized approach to client engagement that FMI Technologies has earned a rich clientele.
That said, the company always remains in the hunt to be at the leading edge of AI innovation by participating actively in some of the most renowned global conferences held by institutions such as Bloomberg, NYU, and Columbia University, among many others. Apart from partnering with institutions such as Princeton, FMI Technologies also interacts with a team of experts who are mostly Ph.D. holders of physics, mathematics, and computer science from institutions such as the University of California, Berkeley, and Stanford, among others. Besides academia, the company maintains strong interactions with people from the investment space to be up to speed with the market needs and developments.
The fact that FMI Technologies has distinguished minds collaborating with it to improve its solutions and services constantly is the major differentiating factor for it in an industry that settles for technicians with lesser expertise. Moving forward, FMI Technologies is heading for a strong collaboration with a big Wall Street service provider in the near future to expand its market reach.