Lead Machine Learning Engineer @ Mindee

Sept. 2023 - Present; Paris, France

As Lead Machine Learning Engineer, I have been developing an AI-based platform for document understanding while leading the ML team to support large-scale inference using LLMs and computer vision. I am revolutionizing our inference stack through QA testing and Kubernetes, optimizing large vector data serving with microservices (pgvector, FastAPI), and enhancing benchmark evaluations. Additionally, I am building Retrieval-Augmented Generation (RAG) systems and developing AI agents for summarization and knowledge distillation.

Senior Machine Learning Engineer @ Jellysmack

April 2022 - August 2023; Paris, France

I supported a team of seven Data and R&D Scientists in deploying computer vision models to production. I was responsible for AI model lifecycle monitoring, building ML libraries, and setting up internal ML frameworks. Additionally, I developed a cloud-based asynchronous Python job to collect 1K assets weekly from an external API and conducted internal audits on MLOps projects and data quality management.

Data Scientist @ FieldBox.ai

Nov. 2019 - March 2022; Paris, France

As a Data Scientist, I led 5+ AI projects in collaboration with clients, ensuring compliance with project constraints. I provided customer support, helping clients adopt AI agents, microservices, cloud, and DevOps technologies. My work spanned industries including oil & gas, rail transport, food, and water technologies.

Data science Research intern @ BearingPoint

April 2019 - Sep. 2019; Paris, France

I worked on time series modeling for manufacturing, deploying prediction models (Prophet, RNNs, SARIMA, RF) using APIs and serverless infrastructure (AWS). I also conducted performance analysis, optimizations, and benchmark evaluations on jupyter notebooks.

Data science Research intern @ Jules Group (Ex Harold Waste)

May 2018 - Sep. 2018; Paris, France

I implemented deep learning algorithms (DBN, CNN, AutoEncoders) for trash image classification. I evaluated model performance, integrated the selected model into a Flask API for mobile app usage, and benchmarked deep learning algorithms against our solution.