Rodrigo Rivera presenting at CDTM

Advancing Machine Learning and TDA for the enterprise


Rodrigo’s research aims to develop machine learning methods that help reduce food waste in emerging markets. He seeks to develop machine learning software that is precise and robust to be deployed by enterprises to optimize supply chains.

To achieve this, he centers his research at the intersection of machine learning and topological data analysis for business problems with societal implications.

Rodrigo has worked with Gaussian processes for demand forecasting and matrix factorization techniques.

Further, he has sought to specialize in topological data analysis for time series. His research has shown him the importance of finding structure in the data. The structure is especially relevant for noisy time series or problems where the aggregation is necessary. A simple example is distributing perishables across a network of supermarkets. The total demand for fresh produce is composed of the sales performed at each point of purchase. As a category, the need for these types of products can be very stable throughout the year. However, at the retail point, it can be very volatile.

Another area of interest is the intersection of machine learning with software engineering. Rodrigo is convinced that machine learning must become engineering science. He is seeking to train graduate students in data science to help them hone their software engineering skills. Through coursework, case studies, and guided practice, the students should write robust machine learning software. Rodrigo’s objective is to prepare them to work in an enterprise environment. He exposes them to working with virtualization, parallelism, testing, and deployment.