Description

I'm a Senior Data Engineer in the Ad Network team at Voodoo and a Committer & PMC member at Apache Airflow.

I have worked on various types of projects, including:

- Building GDPR-compliant Lakehouses and analytics platforms using Apache Iceberg or Apache Hudi.

- Developing low-latency stream applications (stateful and stateless).

- Creating ML platforms on top of Kubernetes clusters, and serving ML models with FastAPI and GRPC.

- Setting up and deploying Spark On Kubernetes clusters with hundreds of jobs and thousands of daily job runs.

Domaines d’expertise

Langues

Français
Bilingue ou natif
Arabe
Bilingue ou natif
Anglais
Capacité professionnelle complète

Préférences en matière de lieu de travail

Accepte de travailler sur site

Châtillon (jusqu’à 50 km)

Apache Airflow
Committer & PMC member
HIGH TECH
avril 2023 - Aujourd'hui (3 ans et 2 mois)
Paris, France
- Active contributor; ﬁx the reported bugs, introduce new features and improve the code quality and its performance.
- Join the discussions and participate in deciding the future of the project.
- Test and vote on the diﬀerent releases, mentor the new contributors and help Airﬂow users to solve their problems.
Airflow Python Kubernetes AWS Helm flask Github Actions GCP Vault SQL
Leboncoin
Senior Data Engineer
E-COMMERCE
octobre 2021 - Aujourd'hui (4 ans et 7 mois)
Paris, France
- Developing low-latency stream applications using Java, Spring Cloud Stream, KStream, Kafka and K8S for fraud detection.
- Develop a scalable in-house feature store to ingest the aggregate and ingest the company events (>1B/day) in a KV store (DynamoDB) using Kafka, Kstream, FastAPI, AsyncIO, Avro and Airﬂow.
- Designing a new Lakehouse architecture to apply GDPR on the legacy datalake and optimize the data processing by optimizing the data ﬁles (compaction, z-order, indexing, ...) using Java Spark, HUDI, Airﬂow, Kafka, S3, Avro, Glue and K8S.
- Improve the data platform: migrate Airﬂow from LocalExecutor to Celery, migrate spark jobs from EMR (YARN) to K8S, migrate Airﬂow operators to the new deferrable (async) mode to reduce the infra cost, migrate Spark to jdk11 after patching the Hive which doesn’t work with jdk11.
- As a Sr. DE, I give data/infra courses (Airﬂow, Spark, Terraform, K8S, ...), I help data teams to design their projects and overcome challenges, and I lead the contribution to open source data projects.
Airflow Spark Python Java Kubernetes AWS Kafka Hudi parquet MLflow Github Actions Terraform avro FastAPI
Data4Risk
Data Engineer & Head of Data
HIGH TECH
février 2019 - octobre 2021 (2 ans et 8 mois)
Paris, France
- Designing and implementing stream applications and batch ETL using Docker, Pyspark, Kafka, Argo Workﬂows, mongoDB, MySQL, MinIO and Kubernetes on GCP and OVH cloud to collect and process weather data.
- Designing and implementing a low-latency Lakehouse using PySpark streaming, Delta Lake, Hive and K8S, to support ACID transactions, update and Delete operations and time travel on the big data tables.
- Leading the DS team: desing and train ML models and pipelines using Keras, Tensorlow, MLlib and other libraries to classify and process satellite images and deploy these models using MLﬂow and TF serving.
- Designing and implementing a datalake for a ﬁnancial data platform: stream Spark on k8s jobs to ingest Kafka events in the parquet datalake, and batch Spark on k8s ETL pipelines scheduled by Argo
Python Spark kafak Delta Lake Kubernetes Argo Workflows MongoDB GCP OVH MLflow Gitlab CI Argo CD Terraform