Chaima YEDES
Verified Expert in Engineering
Databricks Developer
Abu Dhabi, United Arab Emirates
Toptal member since June 17, 2026
Chaima is a data architect with 8 years of experience across data engineering, MLOps, and AI infrastructure. She's a specialist in trusted lakehouses, federated data platforms, and AI-ready foundations. At QData (ADQ), she built a Databricks-native quality framework across multiple entities, and at Valeo, she led the AI data platform for autonomous driving across 4 countries.
Portfolio
Experience
- CI/CD Pipelines - 8 years
- Unity Catalog - 8 years
- Databricks - 8 years
- Python - 8 years
- SQL - 8 years
- PySpark - 8 years
- Apache Airflow - 8 years
- Google Cloud Platform (GCP) - 6 years
Preferred Environment
Azure, Google Cloud Platform (GCP), Collibra, Jenkins, Docker, Terraform, Apache Airflow, Data Build Tool (dbt), Azure Databricks, Snowflake
The most amazing...
...project I've worked on involved architecting a multi-cloud Data Fabric solution across many countries for an autonomous driving AI platform.
Work Experience
Data and AI Architect
Contango
- Architected a governed data marketplace for ADQ holding group.
- Designed federation use cases: cross-company cold warehouse discovery combining storage capacity, supply chain demand, and logistics data in a single federated query, replacing days of phone calls with seconds of querying.
- Built the entire data platform from zero: Databricks lakehouse, Delta Live Tables for CDC, medallion architecture, and data contracts enforced at every producer-consumer boundary.
- Deployed Unity Catalog across multiple ADQ companies in 3 weeks: cataloging, PII tagging, RBAC, column masking, lineage, audit trail.
- Built an automated quality scoring engine reading UC metadata across five dimensions (freshness, completeness, uniqueness, validity, lineage).
- Deployed LLM-powered self-service analytics: business users query governed data in plain English, then the agent generates SQL and returns cited answers with source dataset and freshness score.
Lead Data and MLOps Architect
Valeo (Autonomous Driving Division)
- Inherited 30+ disconnected data sources across Azure, GCP, and on-prem. Designed and built a multi-cloud lakehouse from scratch.
- Architected a Data Fabric approach: one virtual access layer connecting all clouds into a single governed, queryable service.
- Built CDC pipelines from 30+ factory databases: Debezium capturing changes from on-prem MySQL binlogs, synced through Kafka to ADLS.
- Cut a critical ML data pipeline from 8 hours to 35 minutes. Same cluster, 90% cost reduction.
- Implemented three governance tools working together: Collibra Unity Catalog, and Dataplex. Collibra as control plane, UC and Dataplex as enforcement engines.
- Established data contracts between producers and consumers.
- Achieved GDPR compliance on video at scale: YOLOv8 detected faces and license plates in driving footage. All computing was done inside the GCP boundary.
- Enabled multimodal search across 5+ million driving images.
- Deployed ML models via Vertex AI pipelines with CI/CD automation. Leveraged MLflow for experiment tracking, model registry, and versioning, and DVC for dataset versioning pinned to firmware snapshots.
Data Scientist, Engineer Consultant
Demain
- Built ETL/ELT pipelines on Azure (Data Factory, Databricks, ADLS) integrating SQL Server, APIs, and flat files into a governed analytical layer.
- Automated collection and preprocessing. Reduced manual data handling from days to hours.
- Developed pricing anomaly detection on a billion-dollar product catalog. Leveraged DBSCAN for clustering products into pricing neighborhoods, Isolation Forest for scoring anomalies, XGBoost for predicting expected price, and SHAP for explainability.
- Developed a customer propensity-to-buy engine on first-purchase behavioral signals using XGBoost with SMOTE for class imbalance.
- Built customer segmentation and churn prediction models, driving targeted retention strategies across multiple retail clients.
- Created Power BI dashboards for C-suite decision-making.
- Built A/B testing frameworks for campaign ROI measurement.
Junior Data Scientist, Data Engineer
Orange
- Built pipelines to clean, structure, and transform raw telecom signals (connection logs, error counters, Wi-Fi retry rates, system events) into ML-ready datasets at fleet scale.
- Applied DeepSurv (survival analysis) for time-to-failure estimation on right-censored data. Ranked devices into prioritized maintenance queues.
- Performed feature engineering: rolling statistics over multiple time windows, error-burst counts, SNR variance, and firmware version context.
Experience
Data Quality Platform
https://github.com/chaimaYS/llm-data-quality-platformEducation
Master's Degree in Data Science
IMT Atlantique - Brest, France
Master's Degree in Telecommunications
SUP’COM - Tunis, Tunisia
Certifications
Databricks Professional Data Engineer
Databricks
Microsoft Azure Data Fundamentals (DP-900)
Azure
Google Cloud Professional Data Engineer
Skills
Libraries/APIs
XGBoost, PySpark
Tools
Apache Airflow, Cloud Dataflow, Kibana, Collibra, BigQuery, Apache Beam, Terraform, Jenkins, Grafana
Languages
Python, SQL, Snowflake
Frameworks
Delta Live Tables (DLT), Spark
Paradigms
Role-based Access Control (RBAC), DevOps
Platforms
Databricks, Google Cloud Platform (GCP), Vertex AI, Azure, Debezium, Docker, Kubernetes
Storage
PostgreSQL, MongoDB, MySQL, Elasticsearch
Other
Unity Catalog, MLflow, DVC, SQL Server, Delta Lake, Star Schema, Databricks Workflows, CI/CD Pipelines, Looker Studio, Data Science, Data Engineering, Data Governance, Data Quality, Dataplex, YOLOv8, Medallion Architecture, GDPR, Telecommunication Engineering, Security, Large Language Models (LLMs), Kafka, Gemini, FastAPI, DBSCAN, Isolation Forest, Data Build Tool (dbt), Feature Stores, Structured Streaming, Azure Databricks, Computer Vision
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring