Aziz Kastalli
Computer Vision Engineer at Rutilea, Inc.- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
English Professional working proficiency
-
French Professional working proficiency
-
Arabic Native or bilingual proficiency
Topline Score
Bio
Credentials
-
FUNDAMENTALS OF DEEP LEARNING FOR COMPUTER VISION
NVIDIAMar, 2019- Oct, 2024 -
Professional Data Engineer
GoogleDec, 2022- Oct, 2024
Experience
-
Rutilea, Inc.
-
Japan
-
Automation Machinery Manufacturing
-
1 - 100 Employee
-
Computer Vision Engineer
-
Oct 2023 - Present
-
-
Computer Vision Engineer
-
Jan 2023 - Sep 2023
• Detected a person pose using sparse point cloud data by developing a new data-pipeline for the V2V-Posenet model and training it.• Streamlined development and deployment processes by developing and optimizing Docker images.
-
-
-
RUTILEA TUNISIA
-
Tunisia
-
IT Services and IT Consulting
-
1 - 100 Employee
-
Project Lead
-
Jan 2022 - Aug 2022
• Led successful projects which strengthened the clients trust and increased the projects contract. • Broke down projects into tasks, planned schedules and led a team of engineers in an agile environment. • Interacted between Tunisian and Japanese teams on a weekly basis to communicate progress. • Sped up the projects delivery process by developing automated CI/CD pipelines using Jenkins. • Led successful projects which strengthened the clients trust and increased the projects contract. • Broke down projects into tasks, planned schedules and led a team of engineers in an agile environment. • Interacted between Tunisian and Japanese teams on a weekly basis to communicate progress. • Sped up the projects delivery process by developing automated CI/CD pipelines using Jenkins.
-
-
-
Rutilea, Inc.
-
Japan
-
Automation Machinery Manufacturing
-
1 - 100 Employee
-
Computer Vision Engineer
-
Sep 2021 - Dec 2021
• Registered a 3D point cloud scene of a hand moving 3D sensor using Open3D and a custom ICP algorithm. • Assisted Rutilea’s client during a 3 days exhibition and optimized the 3D point cloud registration algorithm’s time execution by 15 times, which made it possible to make quick onsite demos. • Developed an algorithm based on research papers to segment 3D low resolution point cloud. • Registered a 3D point cloud scene of a hand moving 3D sensor using Open3D and a custom ICP algorithm. • Assisted Rutilea’s client during a 3 days exhibition and optimized the 3D point cloud registration algorithm’s time execution by 15 times, which made it possible to make quick onsite demos. • Developed an algorithm based on research papers to segment 3D low resolution point cloud.
-
-
-
GOMYCODE
-
Netherlands
-
Higher Education
-
400 - 500 Employee
-
AI Instructor
-
Aug 2019 - Jul 2021
• Animated workshops about python, machine Learning and deep-learning. • Assisted students during their projects’ development. • Animated workshops about python, machine Learning and deep-learning. • Assisted students during their projects’ development.
-
-
-
TRANSTU
-
Tunisia
-
Transportation/Trucking/Railroad
-
200 - 300 Employee
-
Data Science Intern
-
Nov 2020 - Apr 2021
• Analyzed, cleaned and transformed over 116 million rows of GPS data into a transport timetable CSV file. • Developed an algorithm that rectifies the transport means itineraries to recover over 95% of the data. • Developed a dynamic graph as well as its search algorithm (Uniform-Cost) to solve the earliest arrival and the minimum number of transfer problems within 55 milliseconds on average. • Analyzed, cleaned and transformed over 116 million rows of GPS data into a transport timetable CSV file. • Developed an algorithm that rectifies the transport means itineraries to recover over 95% of the data. • Developed a dynamic graph as well as its search algorithm (Uniform-Cost) to solve the earliest arrival and the minimum number of transfer problems within 55 milliseconds on average.
-
-
-
Factory 619
-
Tunisia
-
Information Technology & Services
-
1 - 100 Employee
-
Data Science Intern
-
Jun 2019 - Jul 2019
• Investigated research papers and collaborated with Factory’s team to propose a solution for detecting babies’ cries. • Extracted and transformed features from the baby cries signal such as zero crossing, signal frequency and bandwidth and traained tree-based and regression models to detect if the signal is a baby cry or not with 83% accuracy. • Applyed Stratified Kfold startegy as well as stacking and ensembling techniques to improve the model's accuracy and reduce bias. • Investigated research papers and collaborated with Factory’s team to propose a solution for detecting babies’ cries. • Extracted and transformed features from the baby cries signal such as zero crossing, signal frequency and bandwidth and traained tree-based and regression models to detect if the signal is a baby cry or not with 83% accuracy. • Applyed Stratified Kfold startegy as well as stacking and ensembling techniques to improve the model's accuracy and reduce bias.
-
-
Education
-
Udacity
Data Engineering Nanodegree, Data Engineering -
Ecole Supérieure Privée d'Ingénierie et de Technologies - ESPRIT
Computer Science Engineering Degree, Data Science -
Marsa Riadh
High School's degree, Mathematics