Aziz Ben Ammar

Key Skills

Software

AWS SageMaker

CVAT

Labelbox

V7 Labs

CloudFactory

Top Subject Matter

Natural Language Processing (NLP)

Computer Vision

Autonomous Systems

Top Data Types

Geospatial Tiled Imagery

Image

Text

Top Task Types

Bounding Box

Entity Ner Classification

Polygon

Segmentation

Freelancer Overview

Experienced AI engineer, excels in TS, NLP, CV, RL, and has a rich background in applied mathematics. Expert in Python, R, C++, using Docker, AWS and Azure to deploy AI solutions. International collaborations and landmark contributions mark my career path. Committed to the advancement of AI, I'm ready to take on new challenges.

Labeling Experience

Rock Type Classification and High-Resolution Geological Imagery Analysis

CVATImageBounding BoxPolygon

Led the development of advanced rock type classification models through the analysis of high-resolution geological imagery. The project involved extensive data labeling efforts, including semantic segmentation, polygon annotation, and bounding box techniques, to accurately identify and classify various rock types in complex geological formations. The labeled datasets were used to train and fine-tune deep learning models, significantly improving the precision of geological surveys and resource exploration. This project set new benchmarks for accuracy and efficiency in geological data analysis.

Led the development of advanced rock type classification models through the analysis of high-resolution geological imagery. The project involved extensive data labeling efforts, including semantic segmentation, polygon annotation, and bounding box techniques, to accurately identify and classify various rock types in complex geological formations. The labeled datasets were used to train and fine-tune deep learning models, significantly improving the precision of geological surveys and resource exploration. This project set new benchmarks for accuracy and efficiency in geological data analysis.

2022 - 2024

Lidar Signal and Image Processing for Maritime Applications

CloudfactoryGeospatial Tiled ImageryGeocoding

Developed and implemented advanced Lidar signal and image processing techniques for maritime applications. The project involved labeling large datasets of Lidar point clouds and images to train models for detecting and classifying underwater and surface objects. Applied point cloud annotation and image segmentation techniques to create detailed training datasets, enabling the development of predictive models for navigation and environmental monitoring. The project also involved applying various signal processing filters like Kalman and Wiener filters to improve the accuracy of Lidar data interpretation.

Developed and implemented advanced Lidar signal and image processing techniques for maritime applications. The project involved labeling large datasets of Lidar point clouds and images to train models for detecting and classifying underwater and surface objects. Applied point cloud annotation and image segmentation techniques to create detailed training datasets, enabling the development of predictive models for navigation and environmental monitoring. The project also involved applying various signal processing filters like Kalman and Wiener filters to improve the accuracy of Lidar data interpretation.

2021 - 2021

Instance Segmentation on Medical Images using Deep Learning Models

LabelboxImagePolygonSegmentation

Developed and implemented instance segmentation models for analyzing 3D medical images, focusing on the detection and segmentation of medical conditions. Utilized deep learning models such as U-Net, Mask-RCNN, and StarDist to create precise masks for various anatomical structures. The project involved extensive labeling of high-resolution medical images, applying polygon annotation and instance segmentation techniques to build a robust training dataset. This work was crucial in advancing the accuracy of automated medical diagnostics.

Developed and implemented instance segmentation models for analyzing 3D medical images, focusing on the detection and segmentation of medical conditions. Utilized deep learning models such as U-Net, Mask-RCNN, and StarDist to create precise masks for various anatomical structures. The project involved extensive labeling of high-resolution medical images, applying polygon annotation and instance segmentation techniques to build a robust training dataset. This work was crucial in advancing the accuracy of automated medical diagnostics.

2020 - 2021

Optimization and Deployment of Video Denoising and Resolution Enhancement Models

LabelboxVideoClassification

Led the optimization and deployment of advanced video denoising and resolution enhancement models, including AutoEncoder and ViDeNN. The project involved labeling video data at the frame level to assess noise characteristics and resolution details, which were used to train and fine-tune the models. Implemented hyperparameter optimization and applied state-of-the-art techniques to enhance the models’ performance, resulting in significant improvements in video quality for real-world applications. Deployed the optimized models on AWS for scalable processing and integration into production workflows.

Led the optimization and deployment of advanced video denoising and resolution enhancement models, including AutoEncoder and ViDeNN. The project involved labeling video data at the frame level to assess noise characteristics and resolution details, which were used to train and fine-tune the models. Implemented hyperparameter optimization and applied state-of-the-art techniques to enhance the models’ performance, resulting in significant improvements in video quality for real-world applications. Deployed the optimized models on AWS for scalable processing and integration into production workflows.

2020 - 2020

Education

Work History