Luc Ebert

#Context: Scrum #Technical Stack: - Python - RAG - AI Agents - OpenShift - Jenkins - Azure - Kubernetes / Docker - Elasticsearch Within a major French banking group, implementation of a strategic AI transformation. The primary objective is to identify and develop a robust and scalable common framework to be utilized across all the company’s entities.

Design and integration of advanced solutions based on generative artificial intelligence. #Context: SAFe #Technical Stack: - Python - Pipenv/Poetry - Angular - Jenkins - Terraform - Kubernetes/Docker - Helm/Flux - MongoDB - PostgreSQL - Elasticsearch - Datadog - Kafka - Sonar - HuggingFace - LiteLLM - Chainlit - Ragas - Pipenv/Poetry - LangChain - Azure OpenAI - Embeddings ● Integration of the LiteLLM solution to securely distribute API keys to internal collaborators in compliance with company standards, thus enabling them to access and benefit from the capabilities of generative artificial intelligence in a standard OpenAI format. ● Implementation of the Chainlit platform, offering a standardized user interface with a pop-in system, facilitating the integration of a customizable chatbot within applications developed by the teams. ● Design of a routing interface between an internal robot dedicated to company-specific knowledge and the GPT model for broader expertise. ● Integration of intent detection to connect GPT with internal and external services through the integration of LangChain agents, thus optimizing the relevance and efficiency of our AI-based interactions. ● Optimization of Embeddings generation on internal documentation by applying diverse strategies, aimed at significantly increasing the accuracy of results. This project highlights my mastery of cutting-edge technologies and generative artificial intelligence, requiring close collaboration with various teams to develop innovative solutions aligned with the strategic objectives of the company.

Development and optimization of a reactive architecture for real-time tracking and monitoring of shipments, ensuring instant and accurate visibility on the status of parcels. Context: SCRUM Technical Stack: - Spark - Scala - Play Framework - React - Cassandra - Kafka - Prometheus - Grafana - Kamon - Elasticsearch - Logstash - Kibana ● Development in Scala with technologies Spark, Kafka, Cassandra, and Elasticsearch ● Expert on Elasticsearch ● Lead on a tracking application processing 10 million events per day (Spark) and called 4 million times per day (Play Framework) ● Agile execution and interactive exchanges with the client ● Anomaly correction ● Performance optimization ● Participation in the drafting of user stories (design and challenge) ● Application deployment This project was a springboard in mastering event-driven and distributed architectures, faced with real-time performance challenges and the management of a considerable volume of events. It strengthened my ability to develop and optimize highly reactive systems, enabling me to master the challenges related to the monitoring and precise tracking of large-scale shipments.

● Contribution to the development of an Exalead application within a nuclear power plant trial digitization project ● Response to Technical and Functional requirements ● Exalead customization in Java ● PLM context with ENOVIA type data indexing

● Development of the UI with Angular 5 for an application for a prestigious school

For an actor in the industrial sector: ● Development of a CATIA macro in VBA to generate tool sheets For an actor in the aerospace industry: ● Development of a CATIA application in VB.NET to import files from AutoCAD software For an actor in the logistics sector: ● Participation in the reorganization of the code for a software solution involving Solr and Banana Internal Achievements: ● Development of a web application with JavaEE and AngularJS interfacing with JIRA ● Development of a Talend script for the Finance department ● Installation of the ENOVIA 2017X solution: installation and configuration of an Oracle database, configuration of Apache HTTP and Apache Tomcat7, creation of security certificates

● Improvement of the signal-to-noise measurement of light intensifiers for the TOPOWL helmet ● Testing light intensifiers on a test bench according to various parameters ● Enhancement of the software enabling measurements on the test bench