Robin Roussel

Robin Roussel

PhD in Computer Science

Smart Geometry Processing Group (UCL)


I have recently finished my PhD in Computer Science at University College London. My advisors were Profs Niloy J. Mitra, Marie-Paule Cani and Jean-Claude Léon. I spent the first year of my PhD at Inria in the IMAGINE team, and 3 months at Adobe as a research intern (mentored by Amanda Ghassaei and Matt Fisher). I was also a postgraduate teaching assistant for 3 years, for the course “Introduction to Machine Learning” (taught by Prof Iasonas Kokkinos). My research primarily focused on computational design methods applied to problems at the intersection of art and engineering. I have published in venues such as Transactions on Graphics (SIGGRAPH) and Computer & Graphics.


  • Computational Design
  • Machine Learning
  • Computer Graphics
  • Computer Vision


  • PhD in Computer Science, 2020

    University College London, UK

  • MSc in Executive Engineering, 2015

    Mines ParisTech, France

  • Scientific Preparatory Class, 2012

    Lycée Masséna, France



Research Intern

Adobe Research

Jun 2019 – Sep 2019 San Francisco, USA

Topic: Generative modeling of chain reaction contraptions

Experimented with deep learning methods to generate shape assemblies, using Python 3 (Jupyter, Numpy, Pytorch) and JavaScript (Three.js).


Postgraduate Teaching Assistant

University College London

Oct 2017 – Jan 2020 London, UK

Course: Introduction to Machine Learning (COMP0088)

Topics: regression & classification, SVM, boosting, neural networks (shallow/deep), PCA, K-means, mixtures of Gaussians

Tools: Python 3 (Jupyter, Matplotlib, Num/Scipy, Pytorch, Scikit-learn)


Research Intern

Dassault Systèmes Corporate Research

Apr 2015 – Sep 2015 Vélizy, France

Topic: Data-driven structure-aware shape synthesis for the semantic reconstruction of indoor scenes from RGB-D data

Wrote an algorithm to generate new 3D models from an existing dataset, in C++ (Eigen, OpenCV, PCL).


Research Intern

University of St Andrews

Jun 2014 – Sep 2014 St Andrews, UK

Topic: Shape reconstruction and body-dimension measurements of cetaceans from 3D sonar scanner data

Created a point cloud processing pipeline (registration, filtering, segmentation, mesh reconstruction), in MatLab & C++ (PCL).


Research Intern

CAOR Mines ParisTech

Sep 2013 – Feb 2014 Paris, France

Topic: Road marking detection and rendering from 3D lidar data acquired with a mobile mapping system

Implemented an image processing pipeline, in C++ (OpenCV).


Designing chain reaction contraptions from causal graphs

Chain reaction contraptions, commonly referred to as Rube Goldberg machines, achieve simple tasks in an intentionally complex fashion …

Exploratory design of mechanical devices with motion constraints

Mechanical devices are ubiquitous in our daily lives, and the motion they are able to transmit is often a critical part of their …

Spirou: constrained exploration for mechanical motion design

Mechanisms are ubiquitous in our daily lives, and the motion they are able to transmit is often a critical part of their function. …