People of ACM - Worawan Diaz Carballo

November 18, 2025

How did you initially become interested in high-performance computing?

In 1998, I attended a presentation on how to make a program run faster by using two computers to work together. It was fascinating and motivated me to take all classes related to parallel computing. Later, when I was pursuing my Master’s degree at the University of Edinburgh, I excelled in high-performance computing architecture taught by my PhD supervisor, Professor Roland N. Ibbett, who was an engineer on the MU 5 machine. 

It was an exclusive learning experience, where we imagined how things were put together decade by decade through the lens of a legacy engineer. My office was in the same building as the Edinburgh Parallel Computing Centre (EPCC). I gained access to their HPCs for my PhD work and received training on using the Cray machine, which I couldn’t have imagined experiencing at Thammasat University in Thailand where I earned my bachelor’s degree. As part of my PhD scholarship commitment, I knew I had to return to Thammasat University to be a lecturer for twice the time I spent on my postgraduate studies. Experiencing research at the University of Edinburgh—a world-class university with extensive resources—I learned and met excellent people. I saw the future me. I witnessed how human power can contribute to a better world when we put effort into utilizing resources efficiently, thereby saving time.

In your most cited paper, “Accelerating Real-Time Face Detection on a Raspberry Pi Telepresence Robot,” you note that effective face detection is essential for achieving motion in telepresence robots, but that this procedure demands high computation power. In the paper, how do you propose addressing this challenge? Can these findings be applied in other scenarios?

The real challenge was—and is—scarcity. The emergence of affordable single-board computers, such as the Raspberry Pi, has lowered the barriers to bringing creativity and optimization to achieve meaningful innovation, especially in many developing countries.  Together with Krit Janard, my student, we sought an affordable way to provide open access to classrooms for those who could not attend in person. That work was before the COVID-19 outbreaks, so remote learning was still unfamiliar. My experience teaching deaf and blind computer science students, combined with related work and budget constraints, led us to choose a pan-tilt robot that could track the speaker's face without requiring the learner’s device to control it. Because the system was designed for classrooms, we integrated VPN-based communication to ensure security and privacy.

We implemented a pipelined, GPU-assisted face detection system that uses the efficient yet low-computational-cost Local Binary Pattern (LBP) algorithm. The system ran two asynchronous threads—one for GPU-accelerated image capture and preprocessing, and another for CPU-based detection and servo control—achieving real-time tracking while maintaining an active video conferencing session. I still remember our triumph when presenting this work in a late afternoon in Spain, where I remotely controlled the robot from across the continent, talking with Krit, who was in the lab late at night in Thailand.

Asynchronous threads, which enable secure face detection and real-time motor control for video communication, are becoming a prominent feature of assistive robotics and are widely adopted in Edge AI innovations. The technological acceleration following the COVID pandemic and the maturity of AI have enabled real-time interpretation and transliteration, making this technology not only more accessible but also slightly more affordable in mobile phone-connected robots.

In a recent ACM blog, you discussed your experiences at the ACM HPC Summer School in Barcelona. What are the main benefits you and your students have gained from the ACM HPC Summer School? How would you like to see the school expand in the future?

The ACM HPC Summer School provided not only in-depth technical training but also a rare opportunity to listen, think, learn, and engage directly with world-class researchers in person. Each school fosters an inclusive, diverse, and intellectually vibrant environment shaped by the unique contributions of participants from across the globe.

For me, the most valuable benefit is the profound sense of renewal it brings—those nostalgic sparks of curiosity that reignite my passion for learning. Attending as both a learner and an educator allows me to regain the technical depth and reflective mindset that can easily fade amid the demands of academic and administrative responsibilities.

For my students, the transformation is even more remarkable. The three who attended the same schools as I did had grown entirely new versions of themselves. Through cross-cultural learning, collaboration, and exposure to global best practices they have come to see themselves as members of a worldwide community of innovators. The technical skills I learned were honed through Thammasat University's student competition teams, enabling me to win international HPC-AI competitions in the APAC region. Winning trophies and strong networking built through HPC seasonal schools earned me trust and opportunities to secure a competitive research grant from Thailand NRCT to develop an HPC capacity-building ecosystem in Northern Thailand. As a result, we onboarded more than 300 newcomers to the HPC world. HPC School experiences have become tangible evidence that striving for opportunities leads to growth far beyond technical skills.

Looking ahead, I hope to see the school expand by strengthening partnerships with ACM Chapters and professional networks to build a sustainable talent development pipeline. Alumni and emerging educators could take on greater leadership roles—forming regional chapters, organizing satellite events, or hosting pre-school workshops to prepare new participants. Additionally, expanding sponsorship and funding opportunities through academic, industrial, and philanthropic partners would help ensure that talented participants from resource-limited backgrounds, including those in rural or developing regions, can attend and succeed.

Such expansions would amplify the school’s transformative impact—creating a multiplier effect that turns one week of intensive upskilling into a long-lasting force for global innovation, collaboration, and institutional reform in HPC education.

Given your research, what is an area within HPC that is poised to make advances in the near future?

Future HPC systems will increasingly combine supercomputers—both traditional and quantum—for modeling, cloud platforms for deployment, and educational frameworks for capacity building, with AI as the backbone. For the past two years, instructors and researchers from the HPC Schools have been sharing visions of the future: AI Scientists, AI Factories, and quantum-classical hybrid jobs on-demand. Results from computations and human interactions with HPCs will become the learning ecosystem itself. It will become an integral part of the training data to improve intelligent agents.

In my opinion, we are entering a sharp corner where AI advancement could either accelerate HPC toward greater efficiency and sustainability, or—and this concerns me—introduce serious new risks. Machine intelligence evolves rapidly and powerfully. However, my experience with onboarding novice users and training competition teams has raised concerns that many HPC users may not be aware of AI's unpredictable responses. AI can make mistakes. I have observed how AI suggestions can waste cycles by creating unnecessary resource blocks in job queues.

This brings urgent questions: How do developing regions adopt AI-accelerated HPC safely? How do we prevent AI optimization from harming novice users? Who ensures equitable access as these systems become more complex? Improving intelligent HPC hybrid workflows in line with ethical frameworks for enhancing human capacity seems likely to be a priority, in my opinion.

The challenge is not just technical but also human: how to communicate and build systems that allocate resources to grow like grass, not pull all of it like a giant tree. To keep the planet earth covered in balance, distributed, resilient, and sustainable. That is where I see the true frontier of HPC.

Why are you working to start an ACM HPC Student Chapter in Thailand?

The seed was planted during my time as a Teaching Assistant at the 2025 ACM ASEAN HPC School in Singapore, when the school director asked, “Why don’t you move the capacity-building activities forward as an ACM Chapter?” That question inspired me to think about sustainability—how to keep the learning momentum alive beyond a single event. Through my HPC Ignite initiative in Northern Thailand, supported by ThaiSC and the National Research Council of Thailand (NRCT), I realized that HPC training builds far more than computational skills; it cultivates systematic, scalable, and performance-aware thinking. These mindsets are essential for tackling real-world challenges such as climate modeling and human digital twins, nurturing climate-conscious innovators who connect technology with sustainability.

In the HPC Ignite project, 373 newcomers from four provinces were trained, and 29 students from high schools and universities proposed 11 HPC-enabled innovations to address local challenges, competing for prizes sponsored by ACM and AWS. One of the most inspiring stories came from a young Hmong participant named Pakon, who developed a real-time Hmong–Thai speech translation system to help his community overcome language barriers in education and healthcare. For two years, he had trained his model on his personal computer without success. With access to the LANTA supercomputer and mentorship through HPC Ignite, his proof-of-concept finally came to life—transforming a personal dream into an impactful innovation now expanding to other ethnic languages. His journey embodies why HPC capacity building matters: it connects local creativity to global infrastructure.

All 29 students received one-year complimentary ACM student memberships, affirming that the global community helps us to stand. Establishing the Thammasat University ACM SIGHPC Student Chapter is the next step in sustaining this movement. The chapter will extend opportunities for students across all universities to access world-class HPC and AI resources, fostering resilience, collaboration, and confidence to engage globally. More than computing, it is about empowering innovation with supercomputers—bringing HPC nodes to local needs, enabling communities to transform creativity into real solutions, and ensuring that every local innovator has a pathway to contribute to the global HPC ecosystem.