Modern approaches to data pipelinesto industrialize your AI applications
Discover how modern MLOps and Data Engineering approaches accelerate the industrialization of advanced AI and GenAI applications. In this session, we explore the state-of-the-art in technologies for guaranteeing the performance, scalability and reliability of AI services.
We'll also be sharing our real-world feedback illustrating the implementation of these solutions and architectures to deploy and maintain AI models at scale, leveraging streaming technologies such as Kafka, inference technologies such as Nvidia Triton and MLOps such as ZenML, ...
Information practical information
When?
🗓️ June 24, 2025, 8:30 a.m. to 10:30 a.m.
Where?
📍 Paris, La Défense, Tour Légende - 20 place de la Défense - 92800 Puteaux
For whom?
🎯 Tech experts, Innovation, CIOs, ... anyone who leads Data/IA teams of any size.
The speakers
Augustin HOFF - Lead Data Scientist at MAIFSpecial guest
This real-life testimonial from MAIF will take you to the heart of the industrialization of their AI services. You'll discover how this leading insurer has architected its streaming data pipelines with Kafka for event-driven management of their solution.
Augustin HOFF, project manager on the Data Science side (MAIF), and Ugo Lorenzini (ILLUIN Technology) will share their feedback and go into detail on the adoption of a microservices approach, demonstrating how it promotes agility, scalability and independent maintenance of the various AI components.
Robert VESOUL
CEO & Co-founder
ILLUIN Technology
Victor ALIBERT
Data Engineering Manager
ILLUIN Technology
Ugo LORENZINI
Staff Data Engineer
ILLUIN Technology
Pierre LECERF
Staff Data Engineer
ILLUIN Technology
Process and themes addressed
8:30 am ☕️ Welcome coffee
9:00 am 🎤 Introduction to ILLUIN Technology
9h15 ⚙️ State of the art in data engineering for automatic re-training of models in production
9:30 am ⚙️ MAIF's experience of industrializing AI services
10h00 🙋♀️ Questions and discussion
ML Ops
By integrating continuous training, deployment and monitoring workflows, MLOps facilitate rapid iteration of models and ensure their robustness in the face of changing data and business needs. This approach is key to moving from isolated POCs to large-scale, resilient and maintainable AI services over time. Through technologies such as ZenML, we will explore how MLOps are becoming the foundation for the industrialization of AI in enterprises.
Data Engineering
The Data Engineering approach is an indispensable link for feeding high-performance AI models, and industrializing AI solutions. Modern data pipelines must meet the challenges of scalability, reliability and governance. By exploring technologies such as Kafka and Airflow, we'll see how Data Engineering can be used to build event-driven data flow architectures and store optimal data.
Why come?
- Discover advanced MLOps / Data Engineering solutions for industrializing your AI models in production
- Benefit from concrete, detailed REX from ILLUIN Technology and its customers, on the challenges of AI industrialization via cutting-edge technologies (Kafka, microservices architectures, Nvidia Triton).
- Identify best practices and pitfalls to avoid for your own AI industrialization projects, based on real-life use cases
- Share your approaches with peers and our experts, and accelerate the production of your own AI models.
Rely on our expertise to turn your data into a strategic asset
We help you design robust architectures and optimize pipelines to guarantee reliable, high-performance data flows. Whether you're just starting out or at an advanced stage, we maximize the value of your data to accelerate the success of your Data projects.
Our expertise
We design ingestion and transformation pipelines for processing massive volumes of data (several million per day). Our skills cover Data Transformation (ETL/ELT pipelines), Batch (Apache Spark...), Streaming (Google Dataflow, Apache Beam / Flink...) and Event (Apache Kafka... ) processing, as well as the orchestration andautomation of complexworkflows using technologies such as Apache Airflow / Dagster.
Examples of technologies used
Our expertise
We centralize and migrate your data into reliable, secure and scalable storage spaces to build your Single Source of Truth (SSoT). Our solutions cover Data Warehouse (Google BigQuery, AWS Redshift) and Data Lake House architectures (on Cloud solutions such as Databricks, or on-premises with a Minio / Dremio / Apache Iceberg stack, for example).
Examples of technologies used
Our expertise
We optimize and deploy your Data infrastructures continuously and automatically (Infrastructure as Code with "Terraform") on your own servers or on the main Cloud Providers (GCP, AWS, Azure, OVH, Outscale...). Your architectures are scalable, secure and fully monitored (metrics/dashboard, alerting, logging, tracing) with a Prometheus / Grafana stack to make your operations more reliable.
Examples of technologies used
Our expertise
We industrialize your AI projects with ML Ops Cloud platforms (AWS Sagemaker / Azure ML / Vertex AI) or on-premise with ZenML, consolidating your pipelines from end to end: from experimentation withexperiment tracking solutions to production deployment, integrating model serving / versioning and orchestrating your training pipelines (with Airflow, for example). Your data is stored, secured and versioned in specialized feature stores for live or batch interrogation.
Examples of technologies used
Our expertise
We offer complete analytics solutions, from processing/storage in OLAP databases ( e.g. Google BigQuery, Apache Druid ) to the integration of data visualization solutions (e.g. Apache Superset, Tableau). Our expertise transforms your data into actionable insights via advanced visualization and reporting tools.
Examples of technologies used
Our expertise
We strengthen the governance, traceability (Data Lineage, Data Monitoring) and security of your data, ensuring its compliance (Data Compliance) and quality throughout its lifecycle. We can help you structure your organization around your data needs, with Data Mesh-type organizations.
Our expertise
We design ingestion and transformation pipelines for processing massive volumes of data (several million per day). Our skills cover Data Transformation (ETL/ELT pipelines), Batch (Apache Spark...), Streaming (Google Dataflow, Apache Beam / Flink...) and Event (Apache Kafka... ) processing, as well as the orchestration andautomation of complexworkflows using technologies such as Apache Airflow / Dagster.
Examples of technologies used
Our expertise
We centralize and migrate your data into reliable, secure and scalable storage spaces to build your Single Source of Truth (SSoT). Our solutions cover Data Warehouse (Google BigQuery, AWS Redshift) and Data Lake House architectures (on Cloud solutions such as Databricks, or on-premises with a Minio / Dremio / Apache Iceberg stack, for example).
Examples of technologies used
Our expertise
We optimize and deploy your Data infrastructures continuously and automatically (Infrastructure as Code with "Terraform") on your own servers or on the main Cloud Providers (GCP, AWS, Azure, OVH, Outscale...). Your architectures are scalable, secure and fully monitored (metrics/dashboard, alerting, logging, tracing) with a Prometheus / Grafana stack to make your operations more reliable.
Examples of technologies used
Our expertise
We industrialize your AI projects with ML Ops Cloud platforms (AWS Sagemaker / Azure ML / Vertex AI) or on-premise with ZenML, consolidating your pipelines from end to end: from experimentation withexperiment tracking solutions to production deployment, integrating model serving / versioning and orchestrating your training pipelines (with Airflow, for example). Your data is stored, secured and versioned in specialized feature stores for live or batch interrogation.
Examples of technologies used
Our expertise
We offer complete analytics solutions, from processing/storage in OLAP databases ( e.g. Google BigQuery, Apache Druid ) to the integration of data visualization solutions (e.g. Apache Superset, Tableau). Our expertise transforms your data into actionable insights via advanced visualization and reporting tools.
Examples of technologies used
Our expertise
We strengthen the governance, traceability (Data Lineage, Data Monitoring) and security of your data, ensuring its compliance (Data Compliance) and quality throughout its lifecycle. We can help you structure your organization around your data needs, with Data Mesh-type organizations.
Some success stories on a large scale
How did Randstad centralize its data with a Data Mart to improve quality and optimize its business processes?
How GEOPOST set up a bespoke, scalable Datalakehouse to boost AI applications?
How has MAIF equipped itself with optimal MLOps infrastructures to scale up its AI projects?
The leaders choose us
To register, click here 👇
🗓️ June 24, 2025, starting at 8:30 a.m.
📍 Face-to-face, Paris, La Défense, Tour Légende - 20 place de la Défense - 92800 Puteaux