Ismaël Zerrouk

As a Business Data Analyst within the B2B Customer Value Marketing (CVM) team at Orange, I play a key role in shaping the data-driven marketing strategy across multiple European countries. My responsibilities include designing and implementing advanced segmentation and scoring models to support acquisition, upsell, and retention campaigns. I contribute to the development of decision-making and performance monitoring tools using technologies such as Tableau/Power BI and Python, with a focus on automation and actionable insights. I also support countries in deploying use cases and structuring universal control groups (UCG) to measure marketing impact. Beyond analytics, I actively drive digital transformation initiatives by improving data governance, aligning marketing with sales and IT, and fostering a test-and-learn culture. My work enables local teams to better understand customer behaviors, prioritize opportunities, and scale successful CVM practices across the Orange Group.

● 🛠️ Designed and developed a predictive analytics tool integrating Python and Power BI for forecasting IT equipment needs. ● 📊 Implemented the Prophet model (Meta) to generate accurate demand forecasts based on historical data. ● ⚙️ Built an automated data pipeline: • SQL queries for data extraction • Python scripts for preprocessing and modeling • Power BI dashboards for interactive visualizations • 🔄 Automated regular prediction updates, reducing manual work and ensuring real-time insights. ● 📉 Impact: • Improved forecasting accuracy by ~25% • Helped optimize stock management and resource allocation • Contributed to an estimated cost reduction of 15–20% in IT equipment procurement and a better RSE (CSR) compliance through smarter reallocation ● 🧠 Gained hands-on experience in predictive modeling, workflow automation, and business-focused data solutions.

During my experience at Institut Teknologi Sepuluh Nopember (ITS), I led the development of machine learning models for processing complaint-related tweets in Indonesian and French. Highlights include: ● 𝐃𝐚𝐭𝐚 𝐇𝐚𝐧𝐝𝐥𝐢𝐧𝐠: Sourced and refined an open-source French tweet dataset using Pandas and Excel. ● 𝐍𝐄𝐑 𝐄𝐱𝐩𝐞𝐫𝐭𝐢𝐬𝐞: Constructed a custom Named Entity Recognition (NER) model using Spacy and Hugging Face Transformers to extract entities like dates and locations from tweets. ● 𝐓𝐰𝐞𝐞𝐭 𝐂𝐥𝐚𝐬𝐬𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧: Tweet classification model using BERT pre-trained model, enhancing categorization of complaints on social media. My experience at ITS honed my skills in NLP, Spacy, Hugging Face, Sklearn, and data augmentation techniques.

-Création d'un MPD (modèle physique de donnée) en SQL pour l'application -Utilisation de Entity Framework (ORM) pour relier le MPD à l'application -Développement de l'application en ASP .NET avec la structure MVC

J'aide et j'accompagne les élèves de primaires de Maurecourt pour leurs devoirs après leurs cours. Cet emploi étudiant me permet de développer beaucoup de compétences RH comme mes capacités d'écoute et de communication.