Onsi Chhima

JADE Project — Finance

Lead the integration of AI in the JADE project, coordinating exchanges with the AI Factory to ensure a smooth implementation. Oversee inter-team collaborations (CQFD, JADE, AI Factory) to improve efficiency and foster data-driven innovation. Coordinate end-to-end deliveries, from design to production deployment. Optimize Databricks environments: cluster configuration, user management, resource optimization, and data governance/security via Unity Catalog. Optimize queries and access on Azure SQL Database (review of joins, filters, indexes, and execution plans) to reduce

DATA QUALITY Project

Implement schema validation rules with JSON Schema Validator to control mandatory fields, data types, specific payroll structures, and source/target compatibility.

Detect anomalies with Python.

Implement monitoring, logging, and error management mechanisms comparable to AWS CloudWatch practices to enhance the observability and operational stability of pipelines.

RDD & HRIS Project

Design and maintain Python/PySpark ETL pipelines orchestrated with Airflow, in a logic similar to AWS Glue processes, to ensure the

transformation, reliability, and scalability of data.

Develop Python/PySpark data transformation pipelines on a Data Lake architecture, utilizing Parquet formats and data stored on S3 to meet analytical and reporting needs.

Process structured and semi-structured files (JSON, CSV, Parquet) by applying PySpark transformations to standardize payroll datasets before integration.

Design decoupled processes relying on event-driven mechanisms (Event Grid / Service Bus), in a logic similar to AWS serverless architectures using Lambda and Step Functions.

Optimize SQL queries used for reconciliation and migration controls, improving the comparison between source payroll data and target HRIS outputs.

Containerize data services with Docker to standardize local development, testing, and deployments across environments.

Apply quality and preprocessing controls on health datasets, including missing value management, consistency checks, and preparation of datasets ready for AI use cases. Contribute to the preparation, processing, and analysis of large health datasets used for AI and machine learning models. Build and maintain data workflows to support research teams in transforming raw and complex data into structured, usable, and reliable data. Automate data pipelines with Airflow to improve repeatability, traceability, and efficiency of processing tasks. Develop data models to streamline processing pipelines in an Azure environment, contributing to a 20% productivity improvement. Prepare cleaned datasets for labeling, training, and predictive analytics use cases, supporting the development of ML models. Containerize data components with Docker and contribute to CI/CD pipelines via Azure DevOps. Apply security-oriented practices, including server-side encryption with Databricks to protect sensitive data.