Antonio Guillen-Perez
AI Research Scientist at Hewlett-Packard Enterprise (HPE)
Email: antonio_algaida [at] hotmail [dot] com
About Me
I am an AI Research Scientist at Hewlett-Packard Enterprise (HPE), specializing in the applied development and deployment of Multi-Agent Deep Reinforcement Learning (MADRL) systems to solve real-world challenges in both Data Center Sustainability and Autonomous Vehicle (AV) optimization. My work has integrated advanced AI techniques to optimize energy efficiency and reduce carbon footprint in large-scale data centers, as well as enhance mobility and safety in urban traffic scenarios for connected autonomous vehicles.
With a Ph.D. in Computer Science, I have translated cutting-edge AI research into impactful applications, including trajectory planning for autonomous driving using imitation learning and reinforcement learning. My research has led to significant reductions in traffic delays and improved coordination at busy intersections.
Autonomous Systems Research
My doctoral work focused on using deep reinforcement learning and imitation learning techniques to generate optimal driving trajectories in multi-agent traffic environments. Key contributions include:
- Intersection Management: Developed algorithms that reduced traffic delays by up to 50% and emissions by 45% in complex urban intersections.
- Learning from Demonstrations: Improved AV behavior by combining expert expertise with multi-agent DRL for safer, more efficient driving strategies.
- Integration with 5G/6G: Leveraged next-generation communication to enable faster model inference and robust vehicle-to-infrastructure connections.
Recent News
- 2024/11: Paper Hierarchical Multi-Agent Framework for Carbon-Efficient Liquid-Cooled Data Center Cluster accepted for the Demonstration Track at AAAI 2025. The paper introduces a hierarchical RL framework for improving the energy efficiency of liquid-cooled data center clusters.
- 2024/10: Paper SustainDC: Benchmarking for Sustainable Data Center Control accepted for presentation at the NeurIPS 2024 Datasets and Benchmarks Track. The paper introduces a novel multi-agent reinforcement learning benchmark for sustainable data center control. Link.
- 2023/12: Received the Best Paper Award at the NeurIPS 2023 Climate Change AI Workshop for work on real-time carbon footprint minimization in data centers using deep reinforcement learning.
- 2022/09: Joined HPE AI Labs as an AI Research Scientist. Initiated and led a pioneering project on Multi-Agent Deep Reinforcement Learning (MADRL) focused on data center sustainability, achieving significant improvements in energy efficiency and carbon footprint reduction.
- 2022/06: Successfully defended my Ph.D. thesis on enhancing the cognitive capability of Intelligent Transportation Systems (ITS) using Artificial Intelligence, awarded Cum Laude. Link.
- 2022/03: Demonstrated a multi-agent AV intersection management system that reduced wait times by 50% in simulation at an international ITS conference.
- 2022/01: Completed my stay at the University of California, Davis as Researcher Visitor. Thanks Prof. J. Sebastian Gomez-Diaz. Focused on AI in healthcare, specifically early detection of throat cancer using a multimodal approach, achieving over 90% accuracy. Link.
Selected Publications
Full publication list at Google
Scholar
Experience
Hewlett-Packard Enterprise (HPE) - AI Labs
Research Scientist, Milpitas-San Jose, CA
- Led multi-agent DRL projects to optimize data center sustainability, achieving >20% reduction in energy consumption.
- Investigated large-scale ML pipelines, including fine-tuning transformer-based models for robustness.
- Collaborated with cross-functional teams: simulation engineers, researchers, and software developers.
Polytechnic University of Cartagena
Associate Lecturer, Cartagena, Spain
- Developed imitation learning and DRL techniques for autonomous vehicles at urban intersections.
- Analyzed 5G/6G communication influence on AV performance and real-time trajectory planning.
- Mentored graduate students in AI, computer science, and robotics.
Dolphin Wave | Startup
ML Engineer and Data Scientist, Murcia, Spain
- Implemented data-driven solutions for client-facing applications, focusing on analytics and modeling.
- Collaborated with a small, agile team to iterate quickly on new ML product features.
Polytechnic University of Cartagena
Research Intern, Department of Information and Communications Technologies, Cartagena, Spain
- Studied UAV-based networking solutions, leading to the publication on Flying Ad Hoc Networks.
- Acquired foundational ML and research skills used later for AV research.
Future Directions
My aim is to further integrate learned trajectory planning and behavior prediction into real autonomous fleets. I am passionate about bridging theory and deployment for safer, more efficient urban mobility. Ultimately, I strive to make robust reinforcement learning solutions a practical reality in next-generation transportation systems.