Computing’s Top 30: Upal Mahbub

When he’s not developing hardware-efficient computer vision solutions for extended reality (XR) or creating novel facial landmark detection algorithms for mobile devices, you might find Upal Mahbub writing poetry in Bengali, creating “zentangle” art, or editing Dorpon, the literary magazine he publishes. Or? You might check the list of Computing’s Top 30 Early Career Professionals for 2024, where you will definitely find him. Mahbub is a staff engineer in the Multimedia R&D Lab at Qualcomm. In the following Q&A, he describes How he combined deep learning with computer vision techniques to develop a real-time hand-tracking framework for XR apps His two-year journey of editing and contributing to a computer vision book that highlights challenges and technology advances in the rapidly evolving field How he revitalized the IEEE Computer Society’s San Diego Chapter by organizing various events, including a popular speaker series that’s now in its third year His current work in 3D reconstruction, which requires algorithms that can efficiently reconstruct 3D scenes from 2D images or video feeds in real-time and can be applied in everything from gaming to remote collaboration What drives his passion for research and writing, both in the science and technology realm and in more creative areas such as children’s literature You received the Early Career Distinguished Alumni Award from the A. James Clark School of Engineering at UMD College Park in 2024. Can you share the key achievements that led to this recognition, and how has your education at UMD shaped your career? I am honored to have received the Early Career Distinguished Alumni Award; this recognition is a testament to the impactful research and innovative contributions I have made in the fields of computer vision, machine learning, and extended reality (XR) technologies. When I first arrived in the U.S. to pursue my PhD at UMD, I was welcomed with open arms. My wife accompanied me, and we started our journey living in the Graduate Hills housing. My PhD advisor, Prof. Rama Chellappa, is an authoritative figure in computer vision and machine learning. I was very fortunate to attend his course and convince him to take me as an advisee. Prof. Chellappa played a crucial role in my development as a researcher and in becoming a better version of myself. He is a true role model, and I have always admired his energy and dedication to numerous research projects, consistently delivering top-notch outcomes. Prof. Chellappa believed in me and encouraged me to visit Google and participate in the Google Advanced Technology and Projects (ATAP) authentication sprint. This was a life-changing event for me, as I had the opportunity to collaborate with researchers from around the world and with Google’s cutting-edge research team in the second year of my PhD. This experience allowed me to continue my collaboration with the Google team, collect data, and publish my research outcomes in multiple papers. Throughout my time at UMD, I had the privilege of learning from and working with several esteemed professors who helped shape my research methodology and thought process. Professors Larry Davis, K.J. Ray Liu, Joseph Jaja, and Min Wu were instrumental in my academic growth. The vibrant and exciting graduate student life at College Park also played a significant role in my personal development. My wife and I enjoyed our walks on campus, visits to the McKeldin Library and Stamps Food Court, and coffee time at the Vigilante Cafe. We moved around and changed homes a couple of times before finally settling near Guilford Street on the second floor of a house. It was a magical time of experiencing life and learning from our surroundings. Overall, my education and life-changing experiences at UMD have been instrumental in shaping my career. The strong emphasis on research excellence, coupled with opportunities for interdisciplinary collaboration has enabled me to make significant contributions to the field of electrical and computer engineering. The support and resources provided by the A. James Clark School of Engineering have been invaluable in my journey, and I am grateful for the recognition of my achievements through this prestigious award. The Outstanding Engineering Service Award from the San Diego County Engineering Council in 2024 is a significant honor. What were the key contributions or projects that led to this award, and how do you approach engineering service and community engagement? Receiving the Outstanding Engineering Service Award was a significant honor and a highlight of my career. This award recognized my efforts and contributions as the Chair of the IEEE Computer Society San Diego Chapter in 2023. Under my leadership, the chapter experienced a revitalization and reached new heights in terms of the quality and breadth of its activities. A key contribution that led to this award was the organization of the 2023 Invited Seminar Series, which featured 11 invited talks from renowned researchers and industry experts from around the world. I was able to convince several distinguished speakers to deliver virtual talks on a wide range of topics, including quantum computing, chaotic systems, geometric estimating, artificial intelligence, synthetic reality, deep neural networks, and neuromorphic computing. Among the notable speakers were Srijan Kumar from Georgia Tech, who discussed advances in AI for web integrity, equity, and well-being; Celia Shahnaz from Bangladesh University of Engineering and Technology (BUET), who presented on 2D biosignal processing for disease detection; and Md Sakib Hasan from the University of Mississippi, who covered neuromorphic computing in two sessions. Other esteemed speakers included Anderson Rocha from UNICAMP, Brazil, who talked about living with synthetic realities created by LLMs and LVMs; Pavan Turaga from ASU, who discussed geometric approaches for robust machine learning; Omar Shehab from IBM, who spoke about quantum computing in the age of ChatGPT; and Atiqur Rahman Ahad from the University of East London, who presented on AI in healthcare. These talks attracted over 1,000 participants globally and were co-hosted by various IEEE societies and chapters, such as IEEE Young Professionals, IEEE Women in Engineering, IEEE Intelligent Transportation Systems, and IEEE Robotics and Automation Society. In addition to the seminar series, I cohosted six more events with various societies and chapters, significantly increasing the total number of events to 16 (compared to just two events in 2022). As a chapter leader, I expanded the IEEE Computer Society executive committee, adding three more vice chairs to help with event organization and a publicity chair to improve the visibility of our events to both members and non-members. The award motivated me to continue with the seminar series in 2024 and now in 2025. In 2024, we hosted 11 talks; three of them were in-person, including presentations from industry experts and a distinguished lecturer from IEEE CS. The number of co-hosted talks increased significantly in 2024, leading to more than 35 events throughout the year. This expansion has allowed us to reach a broader audience and foster greater collaboration within the engineering community. My approach to engineering service and community engagement is rooted in the belief that collaboration and knowledge sharing are essential for advancing the field of engineering. I strive to create platforms where experts can share their insights and experiences, fostering an environment of continuous learning and innovation. By organizing high-quality events and engaging with a diverse group of professionals, I aim to inspire and empower the engineering community to push the boundaries of technology and make meaningful contributions to society. As a staff engineer at Qualcomm Technologies, you work on perception algorithms and systems design for extended reality (XR) use cases. Can you discuss some recent projects you worked on and the innovative techniques you used? At Qualcomm, my work focuses on developing efficient computer vision algorithms for various perception tasks tailored for mobile and XR use cases. These algorithms are optimized for specific hardware accelerators to ensure high performance and low power consumption, which are critical for mobile and XR applications. In the past, I worked on developing a hardware-optimized deep learning solution for facial landmark detection. This project aimed to enhance the accuracy and speed of facial recognition systems on mobile devices. By leveraging deep learning techniques, we were able to create a robust algorithm that can accurately detect facial landmarks even under challenging conditions such as varying lighting and occlusions. Another significant project I contributed to was the development of a real-time hand-tracking framework for XR applications. This framework used a combination of deep learning and traditional computer vision techniques to accurately track hand movements and gestures. My major contribution was to develop a highly efficient deep learning backbone for ego-centric hand and keypoints detection, enabling smooth and responsive interactions in XR environments. Currently, I am working on a project focused on 3D reconstruction, which involves creating 3D models from 2D images or video feeds. This project requires developing algorithms that can efficiently reconstruct 3D scenes in real-time, making it suitable for XR use cases. The 3D reconstruction technology has applications in various fields, including gaming, virtual tours, and remote collaboration. In addition to these projects, I regularly contribute to intellectual property development at Qualcomm. I have been involved in filing several patents related to computer vision and XR technologies. My work often involves tackling new challenges and pushing the boundaries of what is possible with current technology. This continuous pursuit of innovation is what drives me and keeps my work exciting and fulfilling. Reflecting on your career journey—from your education at BUET and UMD to your roles at Qualcomm and other organizations—what are some key lessons you have learned, and how have they shaped your approach to research and innovation? Reflecting on my career journey, I realize how deeply personal and transformative this path has been for me, starting from my upbringing and continuing through my education and professional experiences. I owe a great deal to my early childhood, where I was surrounded by a supportive and inspiring family. I spent 14 years at Ispahani Public School and College (IPSC) in Chittagong, Bangladesh, from kindergarten through 12th grade. My mom taught Bengali literature there, and the teachers of IPSC showered their love and blessings on me. They taught me discipline and encouraged me to pursue higher education. My father, Dr. Mahbubul Haque, who passed away last year, was a charismatic man of many talents. Being a renowned linguist with the second highest national award was just the tip of the iceberg; my father was a freedom fighter, an organizer, a leader, and yet a humble family man. My mom was an equivalent counterpart—the two of them instilled in me a curious mind and taught me to admire every little thing around us. This strong foundation shaped and formed my view of the meaning of life and set the stage for my future endeavors. During my time at BUET, I was incredibly fortunate to have mentors who played pivotal roles in my development. My MS thesis supervisor, Prof. Shaikh Anowarul Fattah, introduced me to research methodology and high-quality research efforts around the world. His guidance was instrumental in shaping my early research career. Additionally, my course teacher and later colleague and collaborator, Prof. Celia Shahnaz, taught me the value of hard work and passion. Both Prof. Fattah and Prof. Shahnaz were key figures in getting me involved with the IEEE Bangladesh section, which laid the foundation for my future leadership roles within IEEE. Another mentor who had a significant impact on my career is Prof. Atiqur Rahman Ahad. He introduced me to traditional computer vision research and got me involved in the ICIEV (and later IVPR) conference. Through him, I had the privilege of gaining first-hand experience in organizing world-class events and later, in publishing edited books. These experiences were invaluable in developing my organizational and collaborative skills. As I noted earlier, my doctoral studies at UMD under the guidance of Prof. Rama Chellappa and other UMD professors were truly transformative. At Qualcomm, I have learned the significance of developing efficient and optimized solutions that can be implemented in real-world applications. Working on projects such as facial landmark detection, hand tracking, and 3D reconstruction has taught me the importance of balancing innovation with practicality. Developing hardware-optimized algorithms that ensure high performance and low power consumption is crucial for mobile and XR applications. This experience has reinforced the need to stay current with technological advancements and continuously push the boundaries of what is possible. A shout out to my Qualcomm and my colleagues and collaborators for their continued support with all my endeavors. Additionally, my involvement in the IEEE community has taught me the value of leadership and service. As the Chair of the IEEE Computer Society San Diego Chapter, I have had the opportunity to organize events and seminars that foster knowledge sharing and professional development. This role has taught me the importance of giving back to the community and inspiring the next generation of engineers and researchers. You have published more than 40 articles and edited two books. What drives your passion for research and writing, and how do you stay motivated to continuously contribute to the scientific community? My passion for research and writing is driven by a sense of duty to the community and to future generations. I believe that sharing knowledge and contributing to the scientific community is essential for progress and innovation. Writing, in particular, is a passion of mine. Beyond scientific writings, I have also published books of poetry and children’s literature in Bengali. This creative outlet allows me to express myself and connect with a broader audience. My wife, Tasnuva Chowdhury, plays a huge role in keeping me motivated. She has been a constant source of support and encouragement since the time we first met. Her belief in my efforts and her unwavering support through every up and down have been invaluable. My family members, including my sister, nieces, and in-laws, also appreciate my efforts and admire my achievements. Their admiration and encouragement push me to strive for excellence and make them proud. I am also driven by a desire to honor the great scholars who have taught me. Mentors like Prof. Shaikh Anowarul Fattah, Prof. Celia Shahnaz, Prof. Atiqur Rahman Ahad, and Prof. Rama Chellappa have had a profound impact on my development as a researcher. Their guidance and belief in my potential have inspired me to push the boundaries of what is possible and contribute meaningfully to the field of electrical and computer engineering. As the Chair of the IEEE Computer Society San Diego Chapter, how do you balance your professional responsibilities with your commitment to community engagement and leadership, and what initiatives are you most proud of? I guess, it requires careful time management and a deep sense of dedication. To manage my professional and IEEE leadership roles effectively, I prioritize tasks, delegate responsibilities when possible, and maintain a clear focus on my goals. Balancing these responsibilities is challenging, but it is incredibly rewarding to see the positive impact of our initiatives on the community. One of the key strategies I use is setting aside dedicated time for IEEE activities. This allows me to plan and organize events without compromising my professional duties. I also rely on a strong team of volunteers and colleagues within IEEE Computer Society who share my passion for community engagement. Their support and collaboration are invaluable in executing our initiatives successfully. I feel most proud that we’ve been able to continue the IEEE CS San Diego’s Invited Seminar Series for the third consecutive year. I couldn’t have done it without tremendous help from a few colleagues from IEEE San Diego Section including Michelle Thompson, Charles Bird, Naveed Kazi, Pirazh Khorramshahi, Mounica Gopisetty, Jessy Ayala, Nazmul Hasan, Nancy Trevellyan, and Terry Hache. Tell us about your doctoral research at the University of Maryland. My doctoral research at UMD was a transformative period in my academic and professional journey, focusing on multi-modal active user authentication for smartphones using computer vision and machine learning techniques. Active authentication involves continuously verifying a user’s identity based on their interactions with the device, rather than relying on a single login event. In early 2015, during the Google ATAP sprint, I developed a facial segment-based face detector. This project aimed to detect partial and occluded facial images captured with a cell phone’s frontal camera under various pose and illumination conditions. Building on this work, I developed improved partial face detection algorithms that outperformed most state-of-the-art face detectors and introduced a novel technique to enhance deep neural networks’ generalization. From 2013 to 2018, as part of the Active Authentication Project funded by DARPA, I coordinated the collection of a large dataset on real-life smartphone usage data. I also developed a model for continuous authentication using location history data from smartphones, which outperformed traditional methods in this domain. This work was recognized with the Best Paper Award at the 2016 IEEE Ubiquitous Computing, Electronics and Mobile Communication Conference (UEMCON) and the Best Poster Award at the 2016 IEEE International Conference on Biometrics: Theory, Applications, and Systems (BTAS). In 2017–2018, I contributed to the IARPA-sponsored JANUS project. During this time, I developed a deep learning architecture for detecting attributes from partially visible faces, which demonstrated excellent cross-domain detection performance. I also developed adversarial networks for fooling high-performance deep classifiers and a real-time face verification framework for home surveillance. Tell us about your experiences editing and working on books—especially the recent book, Computer Vision: Challenges, Trends and Opportunities. What motivated you to work on this project? How did you approach it, and how did you take it to the finish line? Working on Computer Vision: Challenges, Trends, and Opportunities was an incredibly enriching experience that spanned almost two years. The project was motivated by a desire to capture the rapid advancements and diverse applications of computer vision, machine learning, and AI. Collaborating with esteemed co-editors like Md Atiqur Rahman Ahad, Matthew Turk, and Richard Hartley, and co-authoring a chapter with Dr. Vasudev Bhaskaran, provided a unique opportunity to connect with leading researchers and practitioners across the globe. The approach to this project involved several key steps. Initially, we identified the core themes and emerging trends in computer vision that we wanted to cover. This included applications ranging from facial recognition and forensics to medical and rehabilitation technologies, robotics, media, XR, sports, and neuromorphic computing. We reached out to academic leaders and seasoned industrial practitioners to contribute chapters, ensuring a comprehensive and diverse perspective on the field. One of the highlights of this project was the collaboration with Prof. Takeo Kanade, who graciously wrote the foreword for our book. His insights and endorsement added significant value to the work. Throughout the editing process, we maintained regular communication with the contributors, providing feedback and ensuring that each chapter met the high standards we set for the book. Co-authoring chapter 12 with Dr. Vasudev Bhaskaran on the trends and challenges of using XR technologies was particularly rewarding. It allowed us to delve deep into the subject and present a detailed analysis of the current state and future directions of XR. Taking the project to the finish line required meticulous planning and coordination. We set clear deadlines, conducted multiple rounds of reviews, and worked closely with CRC Press to ensure a smooth publication process. The entire journey was a learning experience, from understanding the intricacies of book editing to managing contributions from a diverse group of experts. Working on this book gave me a unique opportunity to collaborate with renowned co-editors and connect with research teams across the world. It was fascinating to learn about cutting-edge work in computer vision, machine learning, and AI, both in academia and industry. The book contains 15 chapters written by experts, covering a wide range of application areas and providing valuable insights for researchers, practitioners, and students. I am also deeply grateful for the support from Qualcomm, which allowed me to contribute to this project; its encouragement and resources were instrumental in enabling me to balance my professional responsibilities with this significant endeavor. You have been working at the forefront of developing an optimized algorithm for XR use cases. Give us some insights into the chapter you wrote in Computer Vision: Challenges, Trends and Opportunities on the challenges, trends, and opportunities in XR. I cowrote the chapter—”Immersive User Experiences: Trends and Challenges of Using XR Technologies”—with Dr. Vasudev Bhaskaran, senior director at Qualcomm. In it, we delve into the fascinating world of XR, which encompasses virtual reality (VR), mixed reality (MR), and augmented reality (AR). XR technologies offer immersive and interactive user experiences, each with unique characteristics. VR immerses users in a completely synthetic environment; AR overlays digital information onto the real world; and MR combines elements of both. Our chapter explores the common perception workloads required for XR devices, which are predominantly computer vision tasks. These include simultaneous localization and mapping (SLAM), feature matching, motion tracking, depth estimation, scene understanding, and virtual avatar tracking. These tasks often need to run concurrently, demanding significant computing resources and resulting in high power consumption. Given the form factor of head-mounted devices (HMDs) used for XR and the limits on heat dissipation, balancing high power consumption with the need for large computing resources is a significant challenge. In the chapter, we discuss how dedicated hardware and distributed computing can help mitigate these challenges, providing an overview of the perception tasks essential for an immersive XR experience. This work highlights the trends, challenges, and opportunities in XR technologies, offering insights into how we can optimize algorithms and hardware to enhance user experiences in this rapidly evolving field. I am grateful to Dr. Bhaskaran for his invaluable contributions and collaboration on this chapter. What interests, skills, or passions do you enjoy apart from technical engineering? Beyond my technical engineering skills, I have a range of interests and passions that keep me engaged. I love photography and often find myself capturing moments and scenes that tell a story. Traveling is another passion of mine, as it allows me to immerse myself in new cultures and environments. Writing is also close to my heart, and I enjoy sharing my thoughts and experiences on my blog. In my free time, I play the guitar, draw zentangles, and occasionally write short poems in Bengali. I’ve had the joy of publishing a poetry book and a children’s book in Bengali. I also enjoy visiting different coffee shops and sharing photos on Google Maps, where my photos have received over 72 million views. Additionally, I edit and publish a literary magazine for short stories called Dorpon (“Mirror” in Bengali). These activities help me express my creativity and connect with others beyond my professional work in engineering and computer vision. Bio: Upal Mahbub Upal Mahbub is a staff engineer in the Multimedia R&D Lab at Qualcomm, San Diego. He received his PhD and MSc degrees in electrical and computer engineering from the University of Maryland College Park, and an MSc in electrical and electronic engineering from Bangladesh University of Engineering and Technology (BUET) in Dhaka. His doctoral research focused on multi-modal active user authentication for smartphones using computer vision and machine learning techniques. Earlier in his career, he was an assistant professor in the Department of EEE at BUET. Mahbub has been involved in impactful research in the fields of computer vision, machine learning, signal processing, and speech enhancement for more than 15 years. His current research focuses on developing hardware-efficient computer vision solutions for XR and mobile applications. At Qualcomm, he has developed hardware-optimized novel Facial Landmark detection algorithms for mobile devices, real-time ego-centric hand pose estimation frameworks for XR, and frameworks to reconstruct 3D scenes from camera feeds for XR use cases. His work on active authentication includes developing facial-segments-based face detectors and improved partial face detection algorithms, which outperform many state-of-the-art face detectors. Mahbub has published more than 40 articles in international conferences and journals, and co-edited books including Contactless Human Activity Analysis (Springer, 2021), and Computer Vision: Challenges, Trends, and Opportunities (CRC Press, 2024). He has served as a guest editor for a special issue in Pattern Recognition Letters, “Advances in Human Action, Activity and Gesture Recognition” and served as an associate editor of International Journal of Computer Vision (IJCV) and program chair of the International Conference on Informatics, Electronics & Vision (ICIEV) and the International Conference on Imaging, Vision & Pattern Recognition (IVPR). He also regularly reviews papers for top CV/ML conferences, including CVPR, ECCV, ICCV, and AAAI, and was recognized as an outstanding reviewer for CVPR 2022. His leadership roles within the IEEE community include serving as the chair of IEEE San Diego Section and the IEEE Computer Society San Diego Chapter, where he has significantly increased the chapter’s activities and engagement. Mahbub’s contributions to the field have been recognized with numerous awards, including the 2024 Early Career Distinguished Alumni Award from UMD’s A. James Clark School of Engineering, and the 2024 Outstanding Engineering Service Award from the San Diego County Engineering Council. He received best paper awards at IEEE UEMCON 2016 and ICCIT 2011, and a best poster award at BTAS 2016. He also received a distinguished graduate fellowship from the A. James Clark School of Engineering at UMD. Mahbub holds nine granted patents, with several more in the filing process, covering a wide range of research topics, including facial landmark estimation, 3D hand pose estimation, always-on camera, and multi-camera tracking systems for AR/VR use cases. His current research on high-performance 3D reconstruction aims to reduce the power consumption and latency of enabling 3D reconstruction on XR devices. He is also collaborating with Prof. Tauhidur Rahman’s lab at the University of California, San Diego, as a Qualcomm Champion on developing event-camera-based systems for pose estimation in low-light, low-power, high-speed scenarios. Mahbub’s dedication to research and innovation, combined with his leadership in professional organizations, makes him a prominent figure in the field of electrical and computer engineering. His work continues to push the boundaries of technology, contributing to the advancement of AI and its applications across various industries. Dig Deeper To learn more about Mahbub and his work, View his publications on Google Scholar Check out his work on GitHub Connect with him on LinkedIn Each week over the next few months, Tech News will highlight different Top 30 honorees. For a full list, see Computing’s Top 30 Early Career Professionals for 2024. In addition to Computing’s Top 30, IEEE Computer Society offers many other awards; to read about the honors and honorees—and nominate the impactful professional in your life—visit the IEEE CS Awards page.