Innovations in the realm of high-performance computing (HPC) are constantly reshaping industries and driving scientific breakthroughs. At the forefront of this technological revolution in Europe stands MeluXina, Luxembourg's state-of-the-art supercomputer powered by LuxProvide. Eric Nilles, DevOps Engineer at LuxProvide, is one of the key minds behind MeluXina. He shares with us his unique journey and insights into the world of HPC as well as the innovative projects he is working on, including the development of the MeluXina cloud platform. 

From hospitality to high-performance computing 

“I began my academic journey with a Bachelor of Business Administration in Hospitality Management in The Hague. My initial focus was on managerial positions within the sector. My various roles in hospitality honed my skills in data analysis and management, particularly thanks to my work in data and revenue management. However, I soon identified a recurring pattern due to my tasks and responsibilities revolving nearly always around data, Business Intelligence (BI) and Excel, even in non-technical roles. This realization prompted a career shift, leading me to pursue a Master's Degree in computer science while working at Cargolux in 2020. I was later offered a position at LuxProvide to serve as user support engineer, marking the starting point of my career in HPC. 

During my time providing user support, I learned a lot from more experienced colleagues allowing me to extend my knowledge and develop a solid foundation in HPC. This environment fostered my growth, leading me to my current position as a DevOps Engineer. In this role, I bridge the gap between development and operations, developing applications but also ensuring their smooth deployment and maintenance. One of my notable projects includes the development of an enterprise resource planning (ERP) tool for managing resources on MeluXina. This tool helps allocate resources efficiently, track usage and provide business intelligence through dashboards.” 

Advancing with MeluXina 

“MeluXina, LuxProvide's flagship supercomputer, offers significant computing power and capabilities. While LuxProvide itself doesn't directly use MeluXina for internal processes, the focus is on making the machine accessible and efficient for customers. The goal is to simplify the use of HPC, making it more user-friendly and accessible to a broader range of users, including those outside traditional HPC fields.  

One of the key initiatives I'm deeply involved in is the launch of a MeluXina cloud. This project aims to extend LuxProvide’s capabilities by offering cloud-based services combining the best of both world: cloud computing’s flexibility coupled to the high-performance of supercomputing. By offering cloud services integrated with HPC capabilities, we aim to provide a seamless experience for users, allowing them to leverage the computational power of the MeluXina supercomputer more flexibly and efficiently. 

MeluXina’s Cloud is designed to facilitate high-performance computing tasks, such as those required for artificial intelligence and large language model applications. This initiative addresses a critical need for scalable, on-demand computational resources that are not only powerful but also readily accessible. We are meticulously building the necessary infrastructure, ensuring it meets nowadays security and operational standards. Our objective is to have a robust, customer-ready cloud service within the year. MeluXina’s cloud services are poised to significantly enhance LuxProvide’s service portfolio, making high-efficiency computing more accessible to a broader range of customers.” 

Greasing the wheels 

“As a DevOps engineer, I act mostly behind the scenes, similar to a machinist, ensuring our customers journey using MeluXina is nothing but smooth sailing. However, keeping up the pace in the ever-evolving industry of technology is quite difficult. Sectors like HPC are really deep and encompass many complex topics and softwares, meaning even after working years in the industry, there is always something new to learn. As such, I strive to continuously broaden my skills and deepen my knowledge in HPC and software development.  

In addition, I work actively towards becoming a certified Kubernetes administrator. This certification will significantly enhance my ability to manage and deploy containerized applications efficiently on our infrastructure. With our upcoming cloud services and the shift towards more containerized workflows, this expertise is becoming increasingly essential to implement the latest best practices and innovations required to provide a seamless experience.” 

Making HPC more accessible 

“I also emphasize the importance of making HPC more accessible and user-friendly. By improving tools and interfaces, we can extend our reach to attract more customers and make the benefits of HPC available to a wider audience. For instance, we integrated the Jupyter Notebook, a web application dedicated to creating and sharing computational documents, to provide our customers a streamlined and custom-made experience tailored to their needs. Collaboration with various organizations and research institutions is key to this effort, ensuring that more entities can leverage the power of MeluXina for their computational needs.”