Nassim Bouberbachene

During this internship, I assisted one of the Odoo development teams at Samartest in the design and implementation of a project management and deliverable history module.

During this final internship, I worked within the Compute Applications team at Kalray. My main tasks were to develop and automate performance tests for CNNs, particularly on acceleration cards and GPUs. My key contributions included: - Maintenance and updating of the KaNN-Models-Zoo, a Git repository containing ready-to-use neural networks, along with supplementary scripts to facilitate their execution on different hardware platforms. - Benchmarking the performance of CNNs on various hardware by automating the execution of algorithms and collecting metrics for performance evaluations. - Implementation of CNN quantization. - Implementation of methods for evaluating the accuracy of classification and object detection models.

During this internship, I carried out a variety of tasks, including analysis, technological research, and the definition of functional specifications for a vehicle tracking and technical advisory application project. My key contributions included: - Study of microservices application architecture style and analysis of existing solutions. - Analysis of user needs to establish a list of key features and microservices. - Definition of communication methods between services and user interactions. - Detailed description of database tables and their entries. - Proposal of suitable technologies and software solutions for developing the different microservices.