Freelancer profile translated to English.
Back to original languageAbout Aziz
As a Machine Learning Engineer / Data Scientist, I am interested in end-to-end data projects, from identifying needs to deploying models into production, using an ML Ops approach.
I have strong skills in:
- Building classification, clustering, or predictive models using a variety of machine/Deep Learning algorithms.
- Deploying models into production in a Cloud solution, such as GCP, or as REST APIs, while closely monitoring the evolution of trained models and associated artifacts.
- Using version control tools like Git for precise tracking of code changes, branch management, and effective collaboration with other development team members.
- Automating ML model deployment and scaling processes using tools like Docker for creating lightweight and portable containers, and Github Actions or Cloud Build for deployment automation.
- Creating pipelines (ETL / ELT) for data ingestion and processing on the GCP platform.
- Creating dashboards with solutions like Tableau or Google Data Studio.
With my experience in diverse fields such as insurance, hydraulics, research, automotive, and luxury, I am ready to help you fully leverage your data and ML models.
French
Native or bilingual
English
Fluent
Remote only
Primarily works remotely
Experience
- L'Oréal SA - L'Oréal FranceSenior MLOps - ML EngineerFASHION AND COSMETICSMay 2021 - Today (5 years and 1 month)Clichy, FranceAs part of this mission, I am responsible for setting up an ML pipeline to train and deploy machine learning models (NLP). These models aim to recommend products by analyzing customer names and descriptions based on their skin characteristics.Tasks performed:- Implementation of MLOps principles to ensure code quality and model performance.- Development of a recommendation system based on perfect match filtering.- Use of Vertex AI pipelines (Kubeflow) for model training and management of associated artifacts (metrics, data, etc.).- Implementation of a CI/CD chain with Cloud Build to ensure code integrity.- Integration of ML models into microservices architectures using tools like Docker and Artifact Registry.- Creation of an API exposing trained models on Cloud Run for easy and quick use.- Participation in setting up infrastructure as code with Terraform for efficient cloud resource management with the DevOps team.- Use of monitoring (Stackdriver) and logging (Cloud Logging) techniques to quickly identify and resolve production issues.Methodology:- Agile (Scrum)Technologies and/or methodologies:Faiss, Pinecone, LLM (Palm2), MLOPS, REST API, Python, NLP, Git, BigQuery, Cloud Build, Kubeflow/Vertex AI, Cloud Run, Artifact Registry, Github, Poetry, Docker, Terraform, Cloud Workflows, VS Code, Machine Learning, Deep Learning
- UmanisCloud Data EngineerSOFTWARE PUBLISHINGFebruary 2021 - May 2021 (3 months)Levallois-Perret, FranceThe goal of this project is to create data processing and ingestion pipelines into the data warehouse (BigQuery).Tasks performed:- Participation in setting up the pipeline architecture to ingest data stored locally into BigQuery.- Transfer of locally stored data to Google Storage (GCS) in a one-time operation.- Ingestion of files stored on GCS into BigQuery using Dataflow to extract specific information from the files before storing them in BigQuery tables.- Orchestration of the different pipeline tasks was carried out using Cloud Composer (Airflow) Python operators.- Creation of dashboards with Data Studio using data stored in BigQuery.
- CNAVData ScientistBANKING AND INSURANCEAugust 2020 - January 2021 (5 months)Paris, FranceDevelopment of a model to detect future claimants at the national old-age insurance fund.• Gathering requirements from business units.• Big Data context and work performed on Cloudera.• Volume: Tens of millions of records * Hundreds of features.• Highly imbalanced dataset.• Use of several algorithms to classify policyholders.• Packaging of developed models.• Facilitating workshops with business units.Technical Environment:• Python, Cloudera CDSW, Spark (pyspark), HadoopMethodology:• Agile (Scrum)
Recommendations
FU
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- MSc - Applied GeophysicsUniversité Pierre et Marie Curie2014- Traitement de signal. - Analyse de données environnementales, physiques, sismique, etc. - modélisation 2D-3D en utilisant des algorithmes d'inversion. - Photogrammétrie. - Scilab-Matlab - Python
- PhD: Implementation of optimization algorithms on physical dataUniversité de Rouen2018Utilisation de données thermiques et géophysiques pour la réalisation de modèles hydrauliques. Ces modèles sont par la suite utilisés pour réaliser des prédictions et de calculs de productivité hydraulique. Outils et Technologies utilisés : Matlab, Python, Comsol, Algorithmes génétiques, Algorithmes hybrides (HMC), Gauss Newton, Metropolis-adjusted Langevin algorithm.
Certifications
- MLOps - Machine Learning operations on AWS and AzureCoursera