During the summer of 2024, between his Bachelor's and Master's studies at Imperial College London, Wadoud Charbak visited the Stochastic Numerics Research Group through The Visiting Research Student Program (VRSP).
During Wadoud's stay, he effectively improved the computational time of a random Fourier features-based Machine Learning training algorithm by writing code in JAX and running it on a GPU.
KAUST: What is your background?
Wadoud: I am currently in my fourth year of the MSci Physics programme at Imperial College London. My studies have focused on both theoretical and computational physics, and covering a wide range of practical aspects and computational aspects around many fields but particularly particle physics.
KAUST: How did you hear about the Stochastic Numerics Group?
Wadoud: I came across the project while browsing the list of internships on the VSRP page, and it immediately caught my eye. The project aligned perfectly with the skills I was eager to develop, particularly in stochastic methods and machine learning. After researching the professor involved, I was deeply impressed by his publications which made me excited to apply and explore these areas further.
KAUST: Why did you choose to come to KAUST?
Wadoud: After speaking with Nawazish, one of the coordinators of the VSRP, I was captivated by KAUST’s remarkable facilities and stunning campus. Its proximity to Jeddah and the chance to live in a culturally rich region, which resonates with my Middle Eastern heritage, were also significant draws for me. KAUST’s reputation as a leading institution in the Middle East, coupled with its cutting-edge research environment, made it the ideal place for me to pursue this internship.
KAUST: What role did KAUST’s resources (e.g., computational facilities, campus environment) play in your research?
Wadoud: KAUST’s computational facilities, particularly the state-of-the-art GPUs, played a pivotal role in my research. The GPUs were essential in the development and execution of the JIT-accelerated algorithm, allowing me to achieve results that would have been impossible otherwise. The combination of these resources and the beautiful, inspiring campus environment created the perfect atmosphere for productive and innovative work.
KAUST: What skills did you develop or improve during your time at KAUST?
Wadoud: My time at KAUST helped me significantly enhance my Python programming skills, particularly with new libraries like JAX. I also became much more proficient in working with Linux systems, especially in optimising GPU usage and I also gained a deeper understanding of machine learning algorithms and developed my ability to critically read and interpret scientific papers—skills that I know will serve me well in the future.
KAUST: Would you recommend that anyone else come to KAUST for a VRSP visit?
Wadoud: Absolutely! The VRSP is a phenomenal experience, offering a combination of cutting-edge research, a beautiful and secure environment, and an incredible cultural setting. The skills and knowledge you gain here are invaluable and will undoubtedly open doors for the future.