Twibot
A bot playing Twinit, a game where we want to match two identical cards. I used a small YOLO model to spot cards, and keypoints matching to match cards.
01
About Me
I am an AI engineer with a strong foundation in computer science engineering, further honed by a specialized year-long focus on artificial intelligence.
My expertise lies in computer vision, particularly in areas such as instance segmentation, object detection, and super-resolution. I am constantly on the lookout for new state-of-the-art models to integrate into both my professional work and personal projects.
Over the years, alongside my technical development, I have cultivated a deep passion for photography, which not only complements my skills but also enriches my understanding of visual aesthetics and image composition.
Take a look to my photographic journey on my Instagram.
Here are a few technologies I've been working with recently:
Languages
Python
C++
SQL
Bash
AI / ML
PyTorch
ONNX
CUDA
scikit-learn
TensorRT
Viz / Proc
OpenCV
scikit-image
Matplotlib
02
Experience
Feb 2022 - present
+
Driving the research and development of X-ray image analysis pipelines for large-scale cargo inspection systems deployed at ports and border crossings worldwide, at Smiths Detection in Paris area.
- ▹ Designed and deployed threat detection pipelines leveraging both classical segmentation algorithms and deep learning models, contributing to state-of-the-art capabilities in X-ray analysis, including the detection of novel threat categories that represent a significant innovation in the field.
- ▹ Developed image quality enhancement solutions using deep learning-based denoising techniques, improving the visual clarity of X-ray scans for operator analysis.
- ▹ Applied instance segmentation methods, from traditional image processing to modern neural architectures, to isolate and classify objects of interest within complex X-ray scenes.
- ▹ Optimized inference pipelines for real-time GPU performance using ONNX and CUDA, reducing latency for production deployment.
- ▹ Co-authored patent filings as part of R&D efforts, reflecting the innovative nature of the work conducted.
PythonPyTorchONNXC++CUDA
April 2021 - Sep 2021
+
I developed deep learning models for damage detection on damaged vehicles using MMDetection library.
- ▹ Writing SOTA report.
- ▹ Implementation of Hybride Task Cascade model.
- ▹ Data management.
- ▹ Training and testing.
PythonPyTorchMMDetection
Feb 2020 - Aug 2020
+
I took part in the development of a Python-based tool allowing migration of RedBack routers’ configuration to Cisco routers.
I carried out several audit on both manufacturers’ routers.
PythonDjango
June 2019 - Aug 2019
+
At Talentsoft I achieved two missions:
- ▹ I was responsible of adding the widget translation of the Talentsoft tool’s landing page.
- ▹ I made a POC to improve the performance of the textual searches performed on the database. I was also responsible of the data migration from MS SQL server to Elasticserch.
C#/.NETTypescriptMS SQL ServerElasticserch
03
Education
2020 - 2021
Master of Science - Artificial Intelligence and Pattern Recognition
Université Toulouse III - Paul SabatierI took one year to specialize myself in Machine Learning and, particularly, in Deep Learning. This program equipped me with solid foundations in both theoretical concepts and practical applications.
I focused extensively on image acquisition and processing techniques.
Additionally, I acquired foundational knowledge in robotics, enhancing my interdisciplinary skills in AI technologies.
2019
Exchange student
Université de Sherbrooke, CanadaDuring one semester, I participated in an exchange program at Université de Sherbrooke, where I acquired foundational knowledge in Data Science and Machine Learning. I also discovered new fields in computer science such as cognitive computing and IoT.
2015 - 2020
Computer Science Engineering
Sup Galilée, Université Sorbonne Paris NordDuring this 5-year program, I acquired a strong foundation in general computer science, including web and software development. The program emphasized significant mathematical components, particularly in algorithm complexity calculation. Additionally, detailed hardware knowledge was gained through courses such as Computer Architecture.
04
Achievements
2-day CUDA course
Foundation of CUDA programming, profiling and debugging
Sogeti Labs Hackaton Jury Prize
Developed ML models to detect people in overindebtedness
05
Projects
Super-resolution methods for brain magnetic resonance images
With my team we implemented three different neural networks to super-resolve brain MRI of marmoset, the world smallest primate.
Scala implementation of UMAP
We contribute to Clustering4Ever, a Scala library, by providing an implementation of Uniform Manifold Approximation and Projection (UMAP).
06
Get In Touch
My inbox is always open. Whether you have a question or just want to say hi, I’ll try my best to get back to you!
Message received — I'll be in touch!
07
