8th of March in Bordeaux
Portraits of colleagues
Meeting with Thalita Firmo Drumond, Cécile Dobrzynski and Marie Martin.
Thalita Firmo Drumond
A PhD student with the Mnemosyne project team, Thalita Firmo Drumond is seeking to improve computer vision and object recognition models, drawing her inspiration from the biological functioning of humans. Her ambition is to develop a digital system that learns faster than those currently used, and which would require fewer images, in order to increase their recognition capacities. Attached to Inria's centre in Bordeaux and to LaBRI (Bordeaux Computer Science Research Laboratory - CNRS, University of Bordeaux), the young Brazilian engineer is physically working at the IMN (Institute of Neurogenerative Diseases - CNRS, Inserm, University of Bordeaux) on the Neurocampus - the 'Silicon Valley of the neurosciences'. Artificial intelligence became a foregone conclusion following the research master's in computer science she dedicated to machine learning. Following a degree in electrical engineering and electronics from the University of Campinas in Sao Paulo (Brazil), Thalita Firmo Drumond tried her hand at embedded systems at the engineering school Télécom ParisTech then as part of an internship at the CEA (French Alternative Energies and Atomic Energy Commission). However her interest in the functioning of the brain and intelligence (as a study subject) caught up with her. Upon her return to Brazil, she identified Inria's work and applied for one of the theses offered by the centre. In 2016, with her application validated, she was delighted to settle in Bordeaux for three years.
at the centre since 2006, Cécile Dobrzynski joined the CARDAMOM project team upon its creation in 2016. Together with three colleagues from Paris, Grenoble and Bordeaux, she is co-developing the Mmg open source meshing platform used by many industry and academic stakeholders. She modifies meshings - these sets of triangles that enable the computer representation of objects - in order to improve their accuracy. This is of great interest to industry players such as Dassault Aviation, who simulate airflow around aircraft in flight conditions.
She also teaches meshing techniques as well as computer programming to engineering students at the engineering school ENSEIRB MATMECA. What Cécile - who rapidly turned towards applied mathematics during her career - likes to do is to develop tools and methods useful for applications.
An engineer with the GeoStat project team, Marie Martin was recruited in July 2017 following a partnership between the project team and the company I2S. GeoStat has developed an expertise and an innovative approach in image processing, in particular in non-convex and sparse optimisation. Marie applies this method to I2S to enable them to optimise performances and to solve colouring, de-noising and reflection removal issues in its I2s DigiBook solution. I2S DigiBook proposes high resolution book scanners, image processing software and online library creation solutions. Marie is very familiar with the Inria Bordeaux - Sud-Ouest research centre since, after her thesis in applied mathematics, she worked as a research and development engineer with the Monc project team for over three years. She contributed to the evaluation and improvement of the performance of an algorithm for the prediction of the growth of pulmonary metastases on clinical cases and to the development of the Nenuphar technology.