Souhail Toumdi

- Mentored and guided learners through Data Science, Machine Learning, and Data Analysis projects. - Assessed and validated deliverables according to professional standards: exploratory data analysis, modeling, optimization, and data visualization. - Provided personalized coaching to help learners achieve their goals, manage their pace, and stay motivated. - Delivered constructive feedback and actionable improvement plans to enhance project quality and learners’ employability.

- Designed and built a candidate recommender system using Elasticsearch, significantly improving the relevance and personalization of job recommendations. - Developed end-to-end pipelines for extracting structured data from resumes by combining Computer Vision (OpenCV, Tesseract) and Natural Language Processing with BERT-based models. - Led multiple R&D initiatives on multilingual word embeddings, as well as job ad and resume representation learning using Siamese neural networks and triplet loss mining to boost matching accuracy. - Deployed machine learning models as scalable microservices, implementing robust CI/CD pipelines with Kubernetes, AWS, and CircleCI. - Collaborated across teams to integrate AI solutions into production, leveraging Kafka for real-time data streaming and GuildAI for experiment tracking.

- Researched and developed state-of-the-art models for Machine Translation, Question Answering, and Language Detection tasks. - Trained and deployed machine learning models for digital marketing use cases in international retail banking, including churn prediction, customer segmentation, and product recommendation. - Designed and implemented an end-to-end client feedback analysis application — combining algorithm development, frontend, and backend — for sentiment analysis and multi-label classification. - Built and administered a GPU server to accelerate deep learning workloads. - Contributed to the creation of a CI/CD pipeline for automated model monitoring and deployment using MLflow. - Supervised an intern on a state-of-the-art survey of Question Answering models, ensuring research quality and technical guidance.

- Study causal inference and its application to investment funds breach detection.