Haithem K.

• Design and build of a highly scalable Performance Management system for giant teleco infrastructures called IntA.

• IntA is a cloud native platform that handles a large volume of data coming from ±100k devices each with ± 10 sub device, logging ±100 raw metrics, reaching 100 millions of input metrics per second.

• Design and implemention a highly scalable data platform with a translation layer, real-time streaming layer and batch analytics layer to enable real-time and large-scale cross-device KQI/KPI calculations.

• Conception of an observability stack with tons of application metrics to monitor the different pipelines and have an operational insight on our solution.

• Collaborating with the test team to build +100 daily integration tests for the different components using behave framework.

• Keywords: Java, Kafka, KStreams, Apache Pinot, Apache Superset, Airflow, Kubernetes, Helm, Behave Framework, Prometheus, Grafana, Jenkins.

• Designing and Implementing a scalable big data platform (PaaS) on Azure cloud, the new platform provides different capabilities to the tenants to schedule, easily scale, monitor, recover and deliver data in different ways (kafka, elasticSearch, Hive tables, FTP).

• Migration of a pillar big data clusters (MapR solution) in Amadeus, handling traffic of 200 data pipelines, 10k jobs/day and more than 30TB/day of data in output.

• Assuring operational tasks on both platform (legacy/ Migrated PaaS on azure cloud) to meet the SLAs contracted with the different clients.

• Keywords: Java, Scala, Spark, Hive, Databricks, Kafka, Airflow, Openshift, Prometheus, Grafana.

◦ Work published in paper: Improving 3D Plankton Image Classification with C3D2 Architecture and

Context Metadata. 10.1007/978-3-030-96299-9 17 2

◦ Study and analysis of digital filters representativity across all the 3D images of Tara plankton dataset.

◦ Transfer learning of different CNN models existing in the literature applied to the study case of plankton spieces.

◦ Keywords: Python, Pytorch, Matplotlib, Pandas, Numpy, Cuda.