Software projects

This page lists some software projects I worked on in the past.

LatNet Builder

LatNet Builder is a C++ open-source software library and tool for constructing highly-uniform point sets for quasi-Monte Carlo and randomized quasi-Monte Carlo methods using state-of-the-art techniques. Quasi-Monte Carlo methods are often used as a replacement for Monte Carlo to integrate multidimensional functions.

LatNet Builder offers various possibilities to use its functionalities:

  • a C++ library
  • a Command Line Tool
  • a Python interface and a Graphical User Interface based on the Jupyter ecosystem
  • a Java interface encapsulated in the Stochastic Simulation in Java (SSJ) software

Test the software online:    

Custom Jupyter widgets

Jupyter interactive widgets enhance the notebook experience by allowing users to create graphical user interfaces. They enable richer interaction with the data and computing resources. I worked with Société Générale and Quantstack to develop custom widgets that build on the framework of the base ipywidgets library in order to connect arbitrary JavaScript libraries to the Jupyter ecosystem.

The project deliverables are meant to be a tutorial for other developers to follow the same path and add their favorite JavaScript library to the Jupyter notebook. To learn more, read our guest post on the Jupyter blog.

Test one of our widgets online:    

Question Answering grounded in Knowledge Graphs

Knowledge Graphs (KG) are a representation of real-world facts, where entities correspond to nodes and subject-predicate-object triples are encoded by labeled edges. Retrieving information in these datastructures is hard, especially due to their size (90 million nodes for Wikidata, 5 billion nodes for Google's KG). In particular, answering factual questions by finding information inside such graphs is a task of interest for applications in voice assistants.

During my internship at Google Research, we worked on two open-source conversational datasets, CSQA and ConvQuestions, both grounded in Wikidata, the KG from the Wikimedia foundation.