Sapiat’s tools and solutions are grounded in our philosophy that unbiased systematic decision-making is central to improving investment performance, and that it is necessary to consistently synthesize a broad array of sometimes inconsistent information sources, from market-derived covariance matrices to soft forecasts from independent research.
With over $100 trillion globally in investable assets managed by asset owners, managers, and other institutions, the stakes are high for getting investment decisions right. Investors need more robust frameworks to explore the impact of uncertainty across different horizons, asset classes, and scenarios. These tools should provide not just mind-numbing numbers in flat presentations— but rather dynamic, interactive, powerful visualizations of the environment and the impact of investment decisions.
Sapiat’s edge is in weaving together the new machine learning methodologies and data science to traditional and alternative data, coupled with an understanding of finance to build tools for investment and risk decisioning, and delivering the solutions built on a contemporary, high performance technology stack.
Traditional quantitative models rely on powerful statistical methods, but often rely on linear approaches applied almost exclusively to market returns. Approaches such as factor modeling are an essential part of the tool kit, but can be fragile to nonlinearities, stress conditions, and long tails.
Markets are not the only source of insight. There are plenty of signals embedded in alternative data sources such as high frequency macroeconomic data, news, and in independent research. However, noise is rampant in these sources also. Information tend to be provided in heterogeneous formats (structured and unstructured), at different sampling frequencies, with different precision and horizons. Synthesizing this information and converting it into actionable “investment intelligence” is at the heart of Sapiat’s mission.
Sapiat believes that the availability of High Performance Computing, large and non-traditional datasets are key drivers in enabling a more systematic and holistic investment process to be applied in mainstream institutional investing. The judicious use of machine inference and learning, coupled with traditional linear techniques (both core competencies of Sapiat) will provide efficient, robust, and more nuanced insights to allocation, portfolio construction, manager selection, and risk.