El Hizaz Moncif

FIRST Team - Hypervision Project: design and implementation of a centralized monitoring system covering Enedis's entire Snowflake fleet (several accounts, hundreds of active pipelines). Objective: ensure service continuity, detect anomalies in real-time, and provide teams with complete visibility into the data platform's health.

• End-to-end monitoring architecture: definition of the Snowflake system metric collection schema (consumed credits, latencies, job failures, query profiles), metric normalization, streaming publication via Kafka deployed on AWS MSK, consumption and storage in Snowflake, real-time restitution on Grafana and Power BI. Fully decoupled and scalable architecture.

• Ingestion zones

Disneyland Paris is currently undergoing a Data transformation by establishing a centralized Data team to serve all company departments and a Data Platform on AWS, enabling DLP to enter a new era of leveraging its rich and varied data from digital, stores, restaurants, hotels, etc. My mission consisted of:

• Feeding the Data Warehouse, Analyzing and Processing source data.

• Setting up the data pipeline: Collect data on our datalake (S3), develop and maintain DP input and output flows on internal and external data, in batch and real-time, up to data activation (Snowflake / Python / Tableau, APIs…)

• Developing and maintaining a Data stack

• Transforming collected data. Enriching and maintaining our Data Warehouse (Snowflake) -> ETL - Ensuring data lifecycle integrity – Designing.

• Setting up and maintaining the entire production pipeline

• Exchanging with our colleagues at The Walt Disney Company in the US, for example, to set up data shares between the US and Paris Snowflake instances

• Proposing, implementing, and supporting best practices within the Data organization (Agile, code review, …)

• Managing data governance issues.

Key Technologies AWS Glue / BATCH / S3 / CodeBuild / Lamda / Snowflake, Python, SQL, Git, SodaSQL, Apache

Disneyland Paris Data Transformation Program: building a centralized Data Platform on AWS from scratch, capable of integrating and unifying heterogeneous data from multiple business domains.

Design and implementation of the complete flow Source -> S3 (Data Lake: landing/raw/curated zones) -> AWS Glue (Apache Spark ETL jobs) -> AWS Batch (heavy processing orchestration) -> Snowflake (Data Warehouse). Clear separation of responsibilities by layer, management of retries and errors at each pipeline stage.

• Batch and near-real-time ingestion patterns: design of reusable ingestion templates (daily batch, hourly micro-batch) in Python/Spark on AWS Glue. Integration of Parquet files from S3 buckets to Snowflake via external stages and COPY INTO commands with schema management.

• Snowflake Data Warehouse: star schema and Data Vault modeling according to business domains, writing ETL transformations (SQL + Python), data lifecycle management, Implementation of Snowflake Data Sharing US - Paris for consolidation of group reporting.

• Infrastructure as Code and CI/CD: provisioning of all AWS resources via Terraform. Automated deployment pipelines with AWS CodeBuild and GitLab CI/CD: unit tests, schema validation.

• Data quality and observability: integration of SodaSQL into pipelines for automatic checks at each load. Monitoring of AWS Glue and Batch jobs via CloudWatch dashboards and SNS alerts. Quality reports available in Tableau.

• Governance and support: definition of the Data Catalog, naming conventions, and code standards. Leading code reviews, implementing Agile methodology.

Technologies: AWS (S3, Glue, Batch, CodeBuild, IAM, CloudWatch, SNS), Snowflake, Apache Spark, Python, SQL, Terraform, SodaSQL, Automic, Docker, GitLab CI/CD, Tableau, Postman, Git