Freelancer profile translated to English.

Description

I support you from start to finish: from data ingestion to its valorization, with a reliable, modern, and impact-oriented approach.

**What I bring you**:

Robust and maintainable data pipelines (Airflow, dbt, BigQuery)

Clear and documented modeling to make your data & business teams autonomous

Useful and actionable dashboards (Looker Studio) to manage your KPIs

Advanced use cases such as content recommendation or intelligent tagging via AI

**My recent missions**:

At Webedia, redesign of ELTs on social media data (Facebook, YouTube, Snapchat) with dbt on GCP: cost reduction, better data quality, budget monitoring, and video tagging via Gemini for audience sales.

Development of a similar house recommendation model integrated into a mobile app and a website, to boost user conversion and enrich the product experience.

**My added value**:

I have a strong ability to simplify complex problems, whether technical or business-related. Curious and adaptable, I work as well with data teams as with product, marketing, or business profiles.

**Types of projects managed**:

Implementation or redesign of modern data stacks (GCP, dbt, Airflow)

Optimization of ELT pipelines (performance, costs, quality)

Deployment of business-oriented dashboards

Integration of simple ML models into the data chain (recommendation, classification)

Languages

French
Native or bilingual
English
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km)

Webedia
Data Engineer
SOCIAL NETWORKS
January 2025 - Today (1 year and 5 months)
Levallois-Perret, France
Optimization & Performance
Implementation of cost optimization strategies on GCP (partitioning, clustering, table expiration) to significantly reduce cloud budget.
Reduction of Dagster DAG processing time through better modularization of dbt models, limiting unnecessary dependencies and segmenting datasets.
Analysis of performance bottlenecks in ELT workflows and targeted refactoring to ensure scalability.
Data Quality & Governance
Definition of data quality rules (presence, uniqueness tests) in dbt to ensure dashboard reliability.
Documentation of data in dbt to enable autonomous use of the provided tables.
Data Valorization & Visualization
Construction of operational and strategic dashboards on Looker Studio for daily monitoring of social media KPIs (engagement, views, budget spent).
Collaboration with social media and marketing teams to identify the right performance indicators.
AI / ML Use Cases
Design of automatic tagging of Facebook content using a Cloud Function that utilizes the Gemini API to classify videos by food themes (e.g., vegan, brunch, snack, etc.).
Integration of generated categories into BigQuery to enable advanced audience segmentation and more targeted inventory sales (data marketing use case).
Stack used
GCP (BigQuery, Cloud Functions, Cloud Scheduler)
SQL, dbt, Dagster
Looker Studio
Python, Git, Gemini API
Python DBT Big Query Dagster Generative AI
Le Collectionist
Data Engineer
REAL ESTATE
April 2024 - January 2025 (9 months)
Paris, France
Data Pipelines & ELT
Design and maintenance of batch ELT pipelines with Apache Airflow, applying Functional Data Engineering principles (simplicity, testability, modularity).
Integration of data sources via Airbyte to feed the Data Warehouse.
Modeling & Transformation
Data modeling in BigQuery in collaboration with Data Analysts to facilitate business exploitation.
Migration of data transformation DAGs from Airflow to dbt, with the aim of:
Reducing operating costs
Simplifying dependency management
Standardizing quality tests (presence, uniqueness, relationship tests)
AI / ML Use Cases
Design and development of a similar house recommendation model, used on the mobile application and website to improve user navigation and increase conversion rates.
Documentation & Autonomy
Implementation of good documentation practices in dbt to make data models accessible to business teams.
Participation in the implementation of a Data Catalog (Secoda) in order to:
Improve data discoverability
Centralize documentation
Facilitate end-user autonomy
Stack used
GCP, BigQuery, Kubernetes, Terraform
Apache Airflow, Airbyte
dbt, SQL, Python
Looker Studio, Secoda, Github
Big Query SQL Python Airflow DBT