Compositional Data Analysis
Multivariate analysis of compositional data with Aitchison’s theory and log-ratio transformations.
Image Quilting of Process-based Geological Models
A computer-vision/geostastical algorithm for conditioning 3D models to data. Useful for very fast geological uncertainty modeling, this software can exploit any modern GPU through the OpenCL standard.
Groundwater Management under Uncertain Seasonal Recharge
A final project for a class at Stanford in collaboration with Thomas Hossler where we model groundwater management with the formalism of Partially Observable Markov Decision Processes (POMDPs) considering the uncertainty of the rainfall in the upcoming year.
Annual Sediment Budget for Stanford Lands
A project for a class at Stanford where I estimate annual erosion on Stanford lands including the campus open space and academic reserve using satellite images.
Education Policy Design in High Schools at Recife, PE
A project in collaboration with Pedro Dantas and Leonardo Rosa where we use Machine Learning to help the city governement of Recife, PE allocate educational resources to students that need the most.
Sequence Mining and Pattern Analysis in Drilling Reports
An internship project at Halliburton Inc. in collaboration with Youli Mao where we classify thousands of drilling reports with Deep Natural Language Processing.
The Inverse Problem of History Matching in Reservoir Engineering
A probabilistic framework for reservoir characterization using high-performance distributed-memory computer clusters.