Chaimae Aboulouafa

Built a RAG evaluation pipeline using Ragas for LLM agent output scoring and benchmarking. Deployed Mistral AI models on Vertex AI (GCP); managed containerized environments with Docker and GitHub Actions. Developed multi-agent systems for data extraction and automated analysis. Fine-tuned Mistral models for ESG domain adaptation, deployed via Ollama. Automated workflows and implemented CI/CD pipelines for reproducible agent testing.

- Improved CI/CD test quality via automated log analysis and error classification using machine learning models. Developed a pipeline to transfer Jenkins CI logs from AWS to GCP. Built unit tests, automated workflows, and analytic dashboards (Looker) Managed data integration across GCP and AWS; applied advanced regex and NLP for error detection Technical Stack: Google Cloud (Vertex AI, BigQuery, Cloud Functions, Cloud Storage,Looker studio, Notebooks, Jenkins, aws S3 bucket

Development of an AI-Driven Social Gaming Platform for Tournament Organization

Designed and launched a profitable social gaming platform where players can create profiles, share posts, join tournaments, and track real-time performance.

Implemented an AI-based recommendation system that suggested tournaments, teammates, and content tailored to each player’s skills and preferences — resulting in a +35% increase in user engagement and +20% tournament participation.

Integrated predictive analytics to forecast tournament outcomes, optimize scheduling, and boost retention of high-value players.

Enabled monetization through targeted in-game advertising and premium tournament subscriptions, increasing revenue potential.

Technical Stack: Node.js, GraphQL, MongoDB, Python, TensorFlow, scikit-learn, OpenAI API.