Freelancer profile translated to English.
Back to original languageAbout Thibault
As a versatile Python developer, my expertise extends well beyond simply developing machine learning models for production. Here's how I can help you:
- Extracting, analyzing, and processing your data, including scraping
- Selecting the most relevant models and algorithms for your project, followed by their modeling and implementation
- Rapid prototyping of solutions to test ideas and iterations
- Developing robust and scalable APIs to expose services to other systems
- Creating efficient backend solutions to manage your business processes
All withcleanandmaintainablePython code that meets industry standards.
I have strong expertise in Python and have worked with a wide range of tools such as: TensorFlow, Keras, Numpy, Pandas, Scikit-Learn, SciPy, Matplotlib, DVC, Poetry, Flask, FastAPI, Django, ...
In addition to being a deep learning engineer specializing in Computer Vision, I am also a competent generalist Python developer. Throughout my career, I have solved a wide range of problems, from real-time region of interest detection for autonomous cars to classification networks for dental pathologies, not to mention reinforcement learning applied to 3D models using OpenGL.
My experience in creating production-ready models, along with my versatility, allows me to adapt quickly, understand, and effectively solve all sorts of problems. I am here to support you in your projects, whether for specific machine learning needs or more general Python development projects.
Feel free to contact me, I look forward to discovering your project!
French
Native or bilingual
English
Fluent
Can work on-site
Paris (up to 50km)
Experience
- Dental MonitoringR&D Engineer in Deep LearningMEDICALDecember 2019 - October 2022 (2 years and 10 months)Paris Area, FranceWithin a company that became aUnicorn**, at the forefront of artificial intelligence in the dental field, I was part of the **deep learningresearch team for almost 3 years.Among the tasks I performed:
- Close collaboration with the business to best determine how to meet their needs,
- Exploration, analysis, and qualification of data necessary for the problem,
- Research and state-of-the-art, followed by the creation and implementation of Deep Learning models,
- Use and optimization of classification and detection models (ResNet, SqueezeNet, YOLO, Mask-RCNN, ...)
- Implementation of models requiring almost infallible predictions as they are linked to medical treatment,
- In-depth analysis and evaluation of model performance before production deployment (Sensitivity, Specificity, Precision, ..),
- Scraping algorithms on databases of several tens of millions of images,
- Management of datasets with rare pathologies and therefore under-represented labels,
- Management of training on datasets of several million images,
- In-depth research in reinforcement learning on topics of 3D model repositioning in space using OpenGL, Gym, and Baselines.
- Use and training of models on AWS servers.
- Constant documentation: project feasibility, progress reports, performance reports, etc.
- Capgemini EngineeringMachine Learning EngineerAUTOMOBILEMay 2019 - November 2019 (6 months)Paris Area, FranceFor 6 months, I was part of the Computer Vision research team applied to autonomous vehicles. Among the tasks performed:* State-of-the-art review of existing visual attention methods and synthesis of the best algorithms
- "Classic" methods using only image processing
- Deep learning methods such as DVA (Deep Visual Attention), DeepFix, DeepGaze, all using an encoder-decoder
* Research, creation, and implementation of these algorithms for detecting areas of importance in the external environment of an autonomous vehicle,- Implementation of the chosen model (DVA here) in Python / Keras
- Adaptation of the model to our needs (dataset, network weights, training duration, metrics, etc.)
* Performance evaluation and model optimization,- Network's ability to generalize, monitoring overfitting/underfitting, performance tracking according to the chosen metric, visual monitoring of performance on real data
- Iterative optimization (loop between optimization and evaluation) of the model using hyperparameters (number of layers, batchnorm, dropout, learning rate, batch size, image normalization, learning rate decay, data augmentation, etc.)
* Adaptation of visual attention methods to a real-time autonomous driving context using Deep Learning.- Network lightening while maintaining a balance between performance and speed
- Use of lighter backbones (VGG16 vs VGG19 for example)
- ECE ParisFinal Project: Autonomous VehicleTECHSeptember 2018 - January 2019 (4 months)Paris, FranceFinal project revolving around the **autonomous car**:The objective was to determine the possibility of exporting an algorithm trained in software (AirSim) to a real-world scaled-down model.A reinforcement learning model to recognize a path and manage obstacle avoidance was trained using Python, TensorFlow, Keras, and reinforcement tools such as Baselines and GYM.
Reviews
Grégoire
Ofeli
Reviewed on 1/3/2023
I hired Thibault for the development of an image classification model.
He put his expertise at my service to guide me in the choice of technologies, feasibility (as well as achievable performance), and finally the end-to-end realization.
The development process was then efficient. Very good regular communication as his work progressed.
Mission completed on time (even a few days early).
I am therefore clearly satisfied with this experience with Thibault and can only recommend hiring him.
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Engineering Degree - Data Sciences & Analytics, specialized in BigDataECE Paris2019* Majeure en systèmes d'information * Spécialisation en Data Science et BigData * Machine Learning et Deep Learning * Réseaux et sécurité informatique * C# * Systèmes d'exploitation
- Neural Networks for Machine Learning (MOOC)Coursera2019Cours dispensé par le professeur Hinton de l'université de Toronto, axé sur l'apprentissage des réseaux de neurones artificiels et leur utilisation pour le Machine Learning, dans le cadre de la reconnaissance de la parole et des objets, de la segmentation des images, de la modélisation du langage et du mouvement humain, etc.
Certifications
- Neural Networks for Machine LearningCoursera Course Certificates2018