Hello 👋

I'm Luca Serra

Freelancer - Senior

ML|

Companies I've worked with:

(More details below ⬇️)

About

➡️ About me

I am a Freelance Senior Machine Learning Engineer currently working for Decathlon Digital.

With 3+ years of experience (3 years as Data Scientist at Artefact), I have developed a strong expertise in deploying machine learning models, forecasting and generative AI, with a foot in computer vision.

I have the strong conviction that every data science project should implement software engineering best practices to make collaboration as smooth and efficient as possible.

👨‍💻 Missions

🛠 My skill set

Mastered languages

PythonSQLBash

Main machine learning expertises

MLOPsGenerative AIForecastingComputer Vision

Daily tech tools

GitDockerGoogle Cloud Platform

Soft skills

Product Management / DeliveryCollaborative Work / Knowledge SharingCommunication / Pedagogy

🏆 Notable achievements

• Open-source contribution to the repository Ultralytics (21k + ⭐️).
• Certified by Google Cloud: Professional Machine Learning Engineer.
• Taught Data Science & MLOPs best practices at top engineering schools (X-HEC, CentraleSupélec, École des Ponts ParisTech, Arts et Métiers).

Projects

Dynamic pricing deployment (2024) Retail Paris

MLOpsAWSDatabricksAirflow

Decathlon Digital

Deployed multiple dynamic pricing projects in Decathlon Digital's infrastructure.

Online Bank Chatbot (2024) Bank Paris

ChatbotGenerative AILLMRAGLangchainFastAPI

BNP Paribas

Built an online bank assistant chatbot using RAG (Retrieval Augmented Generation).

Defect detection - Heat exchangers (2023) Energy Wissous

Computer VisionObject detectionDetectron2Ultralytics

Khimod

Implemented an object detection model to spot defects on metal surfaces of heat exchangers. Contributed to the repository Ultralytics.

Financial forecasting (2023) Beauty Neuilly

ForecastingTime seriesProphetXGBoost

Sephora

Developed a forecasting model to help Sephora to estimate the sales in 5 countries 1 year ahead. Achieved a performance of +10 pts in Forecast Accuracy over current specialists forecasts.

Assortment evaluation (2022 - 2023) Retail Villeneuve-d'Ascq

Assortment optimizationStreamlitGoogle App Engine

Adeo / Leroy Merlin

Built an assortment recommender web application used by 30+ department managers; with a monitoring dashboard.

Forecasting industrialization (2022) Luxury Switzerland

MLOPsForecastingVertex PipelinesCloud SchedulerCloud Functions

Richemont - Montblanc & IWC

Industrialized and automatized two forecasting pipelines using GCP services.

Promotion profitability estimation (2021 - 2022) Retail Ivry

PromotionForecastingProphetXGBoostLooker Studio

Fnac-Darty

Developed a counterfactual forecasting model to estimate past promotions uplift. Link to Medium Article

Promotion uplift estimation POC (2021) Retail Croatia

PromotionProphetLooker Studio

Fortenova

Data science lead on the project to estimate the performance of previous promotion.

Contact

You can contact me on LinkedIn or send me an email if you have any question or want to discuss future opportunities.