CV

Work Experience

Lead engineer and architect of A4S, a national AI trustworthiness sandbox selected as a software pillar of the Luxembourg AI Factory, built in collaboration with major research institutions.
Built a distributed AI testing platform using RabbitMQ + Celery, enabling horizontal scaling and reliable task orchestration.
Automated deployments and server provisioning using Ansible, improving reliability and reducing setup time.
Set up monitoring for compute clusters using Prometheus, Grafana, and Linux-based observability tools.
Designed internal CI/CD pipelines using GitLab CI for reproducible testing and deployment workflows.
Designed and deployed an Ollama- and Docker-based LLM architecture for shared GPU resources, reducing peak memory usage by 3×.
Developed a high-performance experiment runner (Dask + caching), decreasing long-running experimental cycles by 1000×.
Taught AI and machine learning courses at the Master’s level.

Led a 4-year applied research collaboration with BGL BNP Paribas on improving ML robustness for financial transaction systems, resulting in 6 peer-reviewed publications at top-tier venues. See publications
Conducted applied research on adversarial attacks, constrained feature manipulation, and concept drift in high-stakes financial ML pipelines.
Delivered working industrial prototypes to strengthen robustness and monitoring of production-like transaction systems.
Developed TabularBench, an open-source framework for robust evaluation of tabular deep learning; adopted internally for 3 additional projects.
Administered the research group’s computing infrastructure (6 servers, 30 users), automated with Ansible, reducing provisioning time by 5×.
Deployed a secure, self-hosted MLflow platform (Docker Swarm + Ansible), enabling reproducible, privacy-compliant experiment tracking.
Created a lightweight micro-scheduler using Python and Linux pipelines, reducing execution latency by 4× and simplifying batch workflows.
Contributed to internal knowledge-sharing: documentation, reading groups, onboarding sessions.

Developed internal DevOps solutions using GitLab CI/CD, Docker, and Jenkins.
Managed project of a team of 6 student workers using Jira.
Built backend systems in Java Spring Boot to support the university’s student project lifecycle.
Organized advertising and events for the newly created Bachelor in Computer Science (BiCS).

Languages: Python, Java, TypeScript
Data Science & Machine Learning: PyTorch, scikit-learn, HuggingFace, Ollama (GenAI), Pandas, Dask, Celery, Slurm, SQL, Dataiku (Core Designer certification)
Software Engineering: FastAPI, REST APIs, SQLAlchemy, microservices, distributed systems, message-driven architectures, React, MinIO, RabbitMQ
MLOps & Infrastructure: Docker, Kubernetes, MLflow, Ansible, Terraform (learning), AWS (SAA-C03 in progress)
CI/CD & Monitoring: GitHub/GitLab CI, automated deployments, Prometheus, Grafana, observability workflows
Other Tools: Linux, experiment tracking, reproducible pipelines

Thesis: “Enhancing Machine Learning Robustness for Critical Industrial Systems: Constrained Adversarial Attacks and Distribution Drift Solutions.” Read Thesis.
Supervised by Prof. Maxime Cordy. Published 4 papers in NeurIPS, IJCAI, and S&P, including a spotlight at NeurIPS. See publications.

High Honours.
Specialization in Artificial Intelligence and Reliable Software Systems.

High Honours.
Six months exchange at Free University of Bozen-Bolzano, Italy.