leonard poinsot

Working for an AI project aiming at reducing the carbon footprint of malt transformation by optimizing temperature and humidity adjustments. My responsibility is to scale the current product to integrate 40+ factories in the next two years. The main challenge is to integrate and normalize sensor’s data while respecting every production line’s specifics. Technical stack _____________ Azure Databricks, databricks workflows, azure Devops

Assisting strategic consultants in providing excellent data-driven recommendations. (Ponctual weekly mission. So far 2) Technical stack // Power BI - Python (pandas) - SQL Delivering power BI dashboards co-constructed with the strategic analysts for buyer or seller side projects. Working on e-commerce data from various providers.

Architecture, Monitoring and Integration of stores’ traffic data (15 months) ---------------------------------------------------------------------------------- Technical Stack // Python (pandas) - Airflow - Kubernetes - AWS (EKS, Managed Airflow, S3, Vault, Quicksight, Cloudwatch) - Delta - Talend - Opcon - Github Action - Github - Docker - Agile SCRUM Developed an OOP code base allowing to integrate and unit test in 3 days any new API (integration, flattening, normalization, enrichment) Developed a monitoring quicksight and email allowing to know at a glance any issue concerning our product (data quality , user tickets, API) Evolved and monitored a Java API delivering our data to different internal services Integration, Architecture of e-commerce traffic data --------------------------------------------------------- Technical Stack // Python (PySpark) - Airflow - AWS (EMR, Managed Airflow, S3, Vault, Quicksight, Cloudwatch) - GCP (Bigquery, GCS) - Github Action - Github - Docker - Databricks - Agile KANBAN Collected user’s needs and architectured data tables to provide Data Analysts and Analytics Engineer quality data Refined and optimized Airflow workflows by simplifying the code base and the graphs leading to more readability and usability of the workflow Optimized data transformation speed by 40% by optimizing and reading Spark’s DAG to reduce our costs and increase our processing speed.

Creation of a SCV (Single Customer View) at Cora ------------------------------------------------------ Python and PySpark development in a Amazon S3/EC2/EMR environment. The goal of a SCV is to regroup similar customers based on their personal data (name, email etc...). The challenge is to ingest and merge several sources of customers' data, clean and normalize them and as an input and output a single master table regrouping duplicates. The process is automated so that new customers are constantly added to the master data table. Several KPIs were computed following the customers' aggregation. DataLake creation at Zadig&Voltaire ---------------------------------------- Creation of a datahub to regroup most of the companies' data and automate, optimize treatments using a mix of Azure Data Factory and Stored Procedures (as well as Azure services using keyvolt, DevOps with Terraform and C# scripts for Azure Functions...). Data integration, quality and update were the main topics of this mission.

Self-service Dashboard creation for product, technical and commercial teams. The dashboards were made using power BI and allowed a fully automated process (with live connection to PostresSQL DB) for every team to easily dive into data. Data exploration using PBI and Excel to provide deeper insights when needed to each of the teams.

Search engine optimization and content generation for the website. Business development and customer service for our shop-network.