Dr Bei Peng and Amir Ghalamzan have recently joined our School, so we asked them some questions to get to know them:
What are your key areas of research?
BP: My research focuses mainly on deep reinforcement learning, multi-agent systems, and human-in-the-loop machine learning. My long-term research goal is to develop complete intelligent agents that learn how to achieve complex goals from experience, with or without interacting with humans. To achieve this, I work on developing deep multi-agent reinforcement learning methods to enable more effective and scalable coordination among teams of agents.
I investigate human-in-the-loop machine learning algorithms to train agents to learn complex behaviours from human feedback. I also work extensively on applying and extending deep reinforcement learning techniques to other domains, such as crystal structure prediction and residential energy coordination.
AG: My research lies at the intersection of robotic manipulation, learning, and predictive control, with a focus on enabling robots to act safely, intelligently, and adaptively in uncertain real-world environments. I develop data-driven world models that allow robots to anticipate contact events such as slip, deformation, or object motion, and to integrate these predictions into model predictive control for rapid, sub-second decision-making.
A central theme of my work is how robots can combine tactile and visual sensing with machine learning to improve grasping, dexterous manipulation, and interaction with deformable objects. I am particularly interested in developing generalisation mechanisms—through learning from demonstration, transfer learning, and predictive modelling—that allow robotic systems to move beyond narrow task-specific behaviours.
My research is highly application-driven. I have led projects on nuclear robotics, developing autonomous systems for safe sorting and segregation of nuclear waste; agri-robotics, where I pioneered robotic selective harvesting of strawberries using vision and touch; and robotics in healthcare, where I explored early breast-cancer detection using tactile sensing and robotic palpation.
Ultimately, my work aims to establish robots that can predict, plan, and act under uncertainty, creating impactful solutions across manufacturing, agriculture, healthcare, and nuclear decommissioning.
Did something in particular draw you to this research field?
What excites me most is the chance to ensure that robotics and AI are not just drivers of industrial efficiency, but also technologies that bring health and wealth to communities regardless of location or prosperity, enabling fairer opportunities and better futures.
Why did you want to pursue a career in academia and, in particular, in Computer Science?
BP: I enjoy the freedom to work on the research questions I find interesting, and to develop and lead my own research projects. Academia is also the ideal environment to collaborate with many fantastic people from other disciplines and other institutions. I have also found that I really enjoy teaching and working with students. I chose computer science, and specifically artificial intelligence, as my field of study because of the potential for my work to have a significant real-world impact, and the wide variety of problems I am able to work on.
AG: I chose a career in academia because I value the opportunity to educate and inspire the next generation of scientists and thinkers, while advancing research that addresses real-world challenges. Academia provides a unique space to combine teaching, discovery, and collaboration, enabling both immediate impact through students and long-term contributions through research. I pursued Computer Science in particular because it is at the heart of robotics and AI, offering the tools to design intelligent systems that can transform society.
Do you have any recent publications that you would like to highlight?
AG: Yes, I have several that I would like to highlight:
- Nazari and Ghalamzan, “Bioinspired trajectory modulation for effective slip control in robot manipulation”, Nature Machine Intelligence (2025).
- Nazari and Ghalamzan, “Proactive slip control by learned slip model and trajectory adaptation”, Conference on Robot Learning (CoRL) (2024).
- Parsa and Ghalamzan, “Modular autonomous strawberry picking robotic system”, Journal of Field Robotics (2024).
- Sanni and Ghalamzan, “Deep movement primitives: toward breast cancer examination robot”, AAAI Conference on Artificial Intelligence (2022).
What is your favourite thing about teaching the next generation of computer scientists?
What attracted you to working at The University of Sheffield?
BP: Sheffield has a strong reputation in many areas of computer science research relevant to my own interests, including robotics and natural language processing. It also provides excellent support for research, with excellent professional services staff and high-performance computing resources. In my conversations with current staff, I learned that Sheffield attracts extremely strong students at both the undergraduate and postgraduate levels.
Since I began my role here, I have found the School of Computer Science to be a friendly, welcoming, and supportive environment, and I am excited about the many new opportunities for collaboration I have found within the University.
AG: I was attracted to the University of Sheffield by its strong research environment and community in robotics and AI, and the opportunity to contribute to and shape Sheffield Robotics as a collaborative, cross-disciplinary hub. The University’s culture of innovation and collaboration provides an excellent setting to advance cutting-edge research while building meaningful partnerships across academia and industry.
And finally, do you have any hobbies or interests that you would like to share?
BP: I enjoy travelling, particularly in the UK and Europe, and enjoy hiking in the nearby Peak District. I also really like doing Zumba.