Roberto López Castro

PhD Candidate in Computer Science

About

Hi there! My name is Roberto López Castro (Switzerland, 1997). I am currently doing my PhD in Computer Science at Universidade da Coruña (UDC) as a member of the Computer Architecture Group (GAC). My interests cover High Performance Computing (HPC) and Artificial Intelligence (AI), concretely GPGPU programming, sparse computation for Machine Learning (ML) workflows and Automated Machine Learning (AutoML). Some other topics of my interest are computer vision and video processing in real time.

Resume

Education

Teaching assistant

2022 - present

Universidade da CoruñaA Coruña, Galicia

  • Teaching assistant for Integrative Programming. Main content: REGEX, Python, Django and Docker

PhD Candidate in Computer Science

2020 - present

Universidade da CoruñaA Coruña, Galicia

  • FPU fellowship

MSC in High Performance Computing

2019 - 2020

Universidade da CoruñaCoruña, Galicia

  • Grade: 9.78/10
  • 10/10 Final project (with honours)
  • Extraordinary MSc Award

BSc in Computer Science

2015 - 2019

Universidade da CoruñaA Coruña, Galicia

  • (2017 - 2019) Computer Engineering specialization
  • Computer Engineering specialization Grade: 9.2/10
  • 10/10 Final degree project (with honours)
  • Academic Excellence Regional Award, Xunta de Galicia 2019
  • Finalist best open-source final degree project award, CIXUG
  • Second prize best applied final degree project, FIC

Professional Experience

PhD Fellow

Computer Architecture Group (GAC) A Coruña, Galicia
Nov 2020 - present
  • Currently working on my PhD thesis under the support of a FPU fellowship, the most competitive grant from the Spanish Government that funds the development of a PhD thesis during four years.

Visiting Research Fellow @ SPCL group

ETH Zürich Zürich, Switzerland
Sep 2022 - Dec 2022
  • Visiting ETH Zurich (Swiss Federal Institute of Technology Zurich) under the supervision of Professor Torsten Hoefler, Professor of the Scalable Parallel Computing Lab. In this collaboration we focused on Microarchitecture-aware Sparse GPU kernels for Deep Learning.
  • Inditex-UDC Grant for pre-doctoral stays, 2022

Research intern

Computer Architecture Group (GAC) A Coruña, Galicia
Jul 2019 - Oct 2020
  • Hired by the GAC to continue working on HPC and AutoML topics

Collaboration fellowship MECD

Computer Architecture Group (GAC) A Coruña, Galicia
Sept 2018 - Jun 2019
  • Competitive grant from the Spanish Government for initiating outstanding students to research.

Collaboration fellowship UDC

Computer Architecture Group (GAC) A Coruña, Galicia
Oct 2017 - May 2018
  • Competitive grant from the University of A Coruña for initiating outstanding students to research.

Publications

  1. Roberto L. Castro, Diego Andrade. “¿Dónde está Messi? Seguimiento de objetos en vídeos utilizando técnicas de Deep Learning”. In Proceedings of 3rd Machine Learning Workshop Galicia: Vigo, Spain, 2018.

  2. Roberto L. Castro, Diego Andrade. “Using Artificial Vision Techniques for Individual Tracking in Sport Events”. In: Proceedings of 2nd XoveTIC Congress: A Coruña, Spain, 2019. Best communication award

  3. Roberto L. Castro, Diego Andrade. “A Hybrid Approach for Tracking Individual Players in Broadcast Match Videos”. arXiv pre-print, 2020.

  4. Roberto L. Castro, Diego Andrade. “Reusing Trained Layers of Convolutional Neural Networks to Shorten Hyperparameters Tuning Time”. arXiv pre-print, 2020.

  5. Castro, R.L.; Andrade, D.; Fraguela, B.B. OpenCNN: A Winograd Minimal Filtering Algorithm Implementation in CUDA. Mathematics 2021, 9, 2033. https://doi.org/10.3390/math9172033

  6. Castro, R. L., Andrade, D., Fraguela, B. B. Accelerating Machine Learning Computational Kernels on the GPU. 18th International Summer School on Advanced Computer Architecture and Compilation for High-performance Embedded Systems (ACACES), July 13, 2022.

  7. Castro, R. L., Andrade, D., Fraguela, B. B. Probing the Efficacy of Hardware-Aware Weight Pruning to Optimize the SpMM routine on Ampere GPUs. 31st International Conference on Parallel Architectures and Compilation Techniques (PACT 2022), Chicago (USA), October 2022

Professional Service

  1. Reviewer of the scientific journals: Elsevier Expert Systems with Applications.

Contact

Let's get in touch

  • Email: roberto.lopez.castro@udc.es
  • Phone: +34 881 011 376 (extension 1376)
  • Lab: Laboratorio 0.2
  • City: A Coruña, Galicia
  • Affiliation: Computer Architecture Group (GAC)