Alexander Sokol’s Interview on Cloud Computing and ML in Finance
After CompatibL won the bobsguide award for Best Cloud Platform 2020, the company’s Executive Chairman and Head of Quant Research, Alexander Sokol, shed some light on the current technological trends in cloud computing, new approaches to quant modeling, and the machine learning innovations driving digital transformation in the financial industry.
- The cloud in finance is not just for storage anymore
- The pandemic broke barriers to cloud adoption and increased the role of cloud computing in banking and financial services
- Machine learning quant models can prove more effective than traditional ones
- The role of machine learning is becoming more prominent in finance
Cloud Computing is the New Normal for Banking and Financial Services
The cloud today is not just about storage. It is now a completely new paradigm for how applications are built and integrated. Unlike the traditional on-premises architecture, the cloud’s standards-based microservices nature has proved to be real game changer in the digital transformation of the financial sector, leveraging the independency and interoperability of software components.
The Pandemic Accelerated Digital Transformation
One direct impact of the pandemic has been to dramatically accelerate the adoption of many remote working technologies, as the financial industry needed to adapt quickly to rapidly changing market conditions, putting it 5-10 years ahead of its “normal” evolutionary schedule. The role of cloud computing in the financial sector has also drastically increased. Financial institutions have seen at first hand how the cloud provides a faster and more effective way to evolve their software and IT infrastructure than the traditional on-premises model.
The adoption of remote working cloud technologies will likely have a lasting impact when the pandemic ends: financial organizations are unlikely to abandon this powerful ability to manage a geographically distributed workforce and to reduce the need for in-person contact, providing new opportunities for expansion and recruiting.
Machine Learning Modeling Can Succeed Where Traditional Models Failed
Because of the current extreme volatility and uncertainty in markets, there are additional pressures on risk management software to be able to predict the unpredictable. While this is clearly impossible, the new machine learning quant models can definitely help us prepare for what to expect and hedge future risks.
Sophisticated machine learning models used in the financial industry are able to measure both aleatory uncertainty (predictable risk) and epistemic uncertainty (our lack of knowledge about true risk) to identify the boundary between the predictable and the unpredictable, to avoid unwarranted overconfidence in model predictions.
Machine Learning Breakthrough in Finance
We are well past the point where financial institutions were just beginning to investigate the benefits of machine learning. Today, there are many areas in finance that are ready to profit from machine learning upfront. It is a transformative new technology that will touch every aspect of the front, middle, and back office. Machine learning in finance can also outperform traditional, already very sophisticated quant models, such as derivatives valuation, credit risk, and market risk. More importantly, it can fuel a breakthrough in those areas that have traditionally resisted quant math, e.g., operational risk.