Hamza Jeddad

• Architected and deployed multiple specialized AI agents using LangChain and LangGraph frameworks for item classification, item cataloging, valuation, and compliance verification • Developed multimodal solutions integrating text and image analysis apabilities for automated cataloging and compliance checking • Created synthetic training data and fine-tuned Language Models for domain-specific classification and information extraction tasks • Implemented robust monitoring and guardrails with MLflow and dashboards to ensure production reliability • Collaborated with stakeholders to integrate AI solutions into business workflows, ensuring adoption and value delivery Tools: Python, LLMs, LangChain, LangGraph, Vector Databases, Elasticsearch, pgvector, AWS Bedrock, Fine-Tuning, MLflow, Docker, Streamlit, Flask

• Led data-driven measurement of marketing effectiveness, developing methods to quantify campaign impact and optimize marketing investments • Created and adapted machine learning algorithms for the cookieless world (Privacy Sandbox), measuring incremental performance of targeting and retargeting strategies • Developed predictive models for Click Through Rate, View Rate, and Bounce Rate to optimize Real-Time Bidding pricing and media allocation • Implemented A/B testing frameworks to evaluate advertising effectiveness and inform data-driven campaign optimization • Industrialized machine learning pipelines for marketing analytics, ensuring robust model deployment in production • Translated complex analytical insights into actionable recommendations for marketing and business stakeholders Tools: Python, Scikit-learn,Spark, Pandas, PyTorch, Airflow, GCP, GCP Vertex AI, Docker, Looker Studio, Statistical Modeling

• Designed and implemented information retrieval algorithms for textual and visual documents using NLP and computer vision techniques • Developed advanced actuarial machine learning algorithms for New Business Price Optimization • Created and maintained predictive models for Customer Lifetime Value estimation with a 7-year horizon • Developed interactive dashboards using Dash for comprehensive A/B test analysis and KPI monitoring • Presented complex analytical models and business insights to Direct assurance’s executive committee Tools: Python, PySpark, Azure, DataBricks, SQL, NoSQL, Dash / Streamlit

• Implemented BERT Language Model and Transformers for entity recognition in unstructured data • Used fine-tuning and transfer learning techniques to overcome data scarcity and optimize performance Tools: NLP, BERT, Transformers, PyTorch, TensorFlow