Khaled Ben Zina

Architected end-to-end streaming data pipelines through Apache Kafka for retail analytics applications, implementing automated product detection systems for inventory management and operational optimization. Built scalable cloud infrastructure that processed 100M+ daily inputs into ML-ready formats using SQL, Pandas, and BigQuery. Designed and initiated a PySpark project along with Apache Trino and Hive to integrate and transform large volumes of data from company databases into Parquet format, stored in S3 Data Lake for scalable analytics and reporting. Developed automated data workflows for inventory management, creating systems that ensure newly received products are immediately available for analytics and business intelligence pipelines through complex SQL and Pandas-based transformations. Designed and deployed automated orchestration workflows using Apache Airflow (MWAA) as part of a scalable ETL/DataOps framework supporting retail analytics. Collaborated with cross-functional teams to align data architecture with business objectives, delivering robust and cost-effective solutions that improved operational efficiency.

Migrated the company's IT infrastructure to AWS (EC2, S3, Lambda, SQS, Redshift, Postgres, OpenSeacrh), improving scalability, reliability, and achieving significant cost savings. Architected and deployed serverless AWS Managed Workflows for Apache Airflow (MWAA) infrastructure from scratch in a production environment, designing a fault-tolerant, scalable workflow orchestration system with custom operators and sensors that integrated seamlessly with existing AWS services (S3, Lambda, EC2) and reduced pipeline failure rates by 40%. Developed Pandas/SQL-based ETL pipelines to automate data ingestion, transformation, and loading, improving processing efficiency and ensuring data accuracy. Led Data Warehouse migration from Redshift to AWS Athena and S3, implementing a Data Lake architecture, resulting in $30K in annual savings while improving data accessibility, query performance, and scalability for analytics and reporting.

Designed and developed a Deep Learning NLP algorithm in Python to classify 100M+ maintenance task data lines, significantly improving operational efficiency and classification accuracy. Led the integration of data from 5 newairline companies into Airbus's database, overseeing data cleaning, analysis, and collaborating with clients to ensure data accuracy and consistency. Collaborated with Palantir Technologies on implementing a large-scale big data environment for Airbus's global maintenance operations, resulting in a 30% reduction in aircraft downtime through predictive analytics. Developed interactive Spotfire dashboards and analytical applications for C-level executives, providing real-time visibility into fleet maintenance metrics and KPIs across 4 continents. Architected SQL database schema optimizations and stored procedures that improved query performance by 45% across critical maintenance reporting systems. Analyzed complex datasets to identify anomalies, trends, and risks, delivering actionable insights that optimized internal controls and improved maintenance operations.