Maria LEBEDEVA
Data Scientist Computer Vision at TORUS AI- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
Russe Native or bilingual proficiency
-
Anglais Professional working proficiency
-
Français Professional working proficiency
Topline Score
Bio
Credentials
-
TOEIC : 975/990
ETS GlobalJan, 2021- Nov, 2024 -
TOEFL : 100/120
ETS GlobalFeb, 2020- Nov, 2024 -
Program for filling missing data with the use of machine learning methods
Федеральный исследовательский центр «Информатика и управление» Российской Академии НаукNov, 2019- Nov, 2024 -
Application software packages for statistical data analysis - SAS Base, SAS Macro, SAS STAT
SASFeb, 2019- Nov, 2024
Experience
-
TORUS AI
-
France
-
IT Services and IT Consulting
-
1 - 100 Employee
-
Data Scientist Computer Vision
-
Jun 2023 - Present
-
-
-
ISIR - Institut des Systèmes Intelligents et de Robotique
-
France
-
Research Services
-
1 - 100 Employee
-
Research Engineer Natural Language Processing
-
Dec 2021 - Jan 2023
Research subject : New Deep Learning Methods for Data-to-Text and Text-to-Data Generation - built a dataset corresponding to the task and developed a language model - trained large language models to learn from unaligned corpora in a cycle framework - explored new formalisms for learning mappings from diverse sources - focused on controlled text and data generation, according to different aspects and user needs Research subject : New Deep Learning Methods for Data-to-Text and Text-to-Data Generation - built a dataset corresponding to the task and developed a language model - trained large language models to learn from unaligned corpora in a cycle framework - explored new formalisms for learning mappings from diverse sources - focused on controlled text and data generation, according to different aspects and user needs
-
-
-
-
Research Intern
-
Apr 2021 - Sep 2021
Internship subject : High Performance Generative Adversarial Network (GAN) for Optimized Monte Carlo Simulations - developed and trained GANs associated with a Monte Carlo-based simulator - implemented different types of GANs on CPU and GPU architectures using PyTorch library - obtained a model that generates trajectories as accurate as a large Monte Carlo simulation but faster - applied the developed model to the Backward Stochastic Differential Equations Internship subject : High Performance Generative Adversarial Network (GAN) for Optimized Monte Carlo Simulations - developed and trained GANs associated with a Monte Carlo-based simulator - implemented different types of GANs on CPU and GPU architectures using PyTorch library - obtained a model that generates trajectories as accurate as a large Monte Carlo simulation but faster - applied the developed model to the Backward Stochastic Differential Equations
-
-
Education
-
Sorbonne Université
Master 2 (M2), Mathématiques et Applications -
Institut de Statistique de l'Université de Paris - ISUP
Master 2 (M2), Ingénierie Statistique et Data Science (ISDS) -
Московский Государственный Университет им. М.В. Ломоносова (МГУ)
Master of Science - MS, Applied Mathematics and Computer Science -
Санкт-Петербургский Государственный Университет
Bachelor of Science - BS, Mathematical and Statistical Methods in Economics