Francis Williams is a Senior Research Scientist at NVIDIA with 12 years of experience applying 3D deep learning and geometry processing to real-world problems. He combines a strong research background (PhD work at NYU) with hands-on systems and C++ engineering, contributing to both production-focused libraries and research prototypes. His open-source work includes point-cloud utilities and contributions to keops and libigl—adding mathematical primitives, Python bindings, and efficient nearest-neighbor and sampling routines used in geometry and kernel computations. Past roles at Google Brain, MemSQL, and multiple internships show a pattern of tackling graphics, distributed systems, and AR—moving fluidly between low-level performance work and higher-level research. Based in New York, he has repeatedly delivered practical implementations of complex algorithms (e.g., signed distance functions, Poisson disk sampling, atan2 derivatives) that bridge theory and deployable tooling. Colleagues can expect a researcher-engineer who writes production-quality C++ while pushing the state of 3D ML and geometry libraries.
12 years of coding experience
4 years of employment as a software developer
Doctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at New York University
High School Diploma High School/Secondary Diplomas and Certificates, High School Diploma High School/Secondary Diplomas and Certificates at Stanstead College
Bachelor of Applied Science Honors Software Engineering, Bachelor of Applied Science Honors Software Engineering at University of Waterloo
Contributions:84 releases, 7 reviews, 455 commits in 4 years 4 months
Contributions summary:Francis's primary contributions involve implementing and integrating features related to processing and manipulating 3D point clouds and meshes. This includes the development of C++ code for tasks such as Poisson disk sampling, clustering, and random sampling of mesh vertices, as well as providing a function for calculating the signed distance to the mesh. The commits demonstrate the integration of nanoflann for nearest neighbor queries and the creation of utility functions for I/O operations on mesh files, further enhancing the library's functionality. The user added new functions for processing point clouds.
Simple MPL-2.0-licensed C++ geometry processing library.
Role in this project:
Back-end Developer
Contributions:32 commits, 21 PRs, 13 pushes in 2 years 3 months
Contributions summary:Francis contributed to the Python bindings for the `libigl` library, focusing on `igl_adjacency_list` functionality and fixing bugs within the `generate_docstrings.py` script. This included adapting the code for the python bindings, adding support for an adjacency list, and addressing issues related to argument handling and clang compatibility. The user also modified the python modules and typedefs.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.