NUS Launches MAGIC Lab to Advance Robotics and AI Research in Singapore

The National University of Singapore (NUS) School of Computing has officially launched the MAGIC Lab, marking a significant milestone in Singapore's higher education landscape for advanced robotics and artificial intelligence research. Established in 2026, this new facility is dedicated to developing robotic intelligence capable of perceiving, reasoning, and acting effectively in complex physical environments. The lab's creation aligns with Singapore's ambitious National AI Strategy 2.0, which includes over S$1 billion in public investments from 2025 to 2030 to bolster AI research capabilities and position the nation as a global AI hub.

MAGIC Lab, short for Manipulation and General Intelligence Control Lab, addresses critical challenges in embodied AI, where robots must interact seamlessly with the real world. This initiative comes at a pivotal time as Singapore's universities ramp up efforts to translate academic breakthroughs into industrial applications, particularly in manufacturing, healthcare, and logistics. NUS, consistently ranked among the top 10 universities worldwide by QS World University Rankings, is leveraging its strengths in computing and engineering to lead this charge.

The launch underscores NUS's commitment to fostering interdisciplinary collaboration. By integrating machine learning, 3D vision, and dexterous manipulation, the lab aims to create deployable robotic systems that deliver tangible economic and societal value. Early projects already demonstrate promise, setting the stage for transformative impacts on Singapore's tech ecosystem.

Leadership: Assistant Professor Jiafei Duan's Vision

At the helm of MAGIC Lab is Assistant Professor Jiafei Duan, an incoming Presidential Young Professor at NUS School of Computing. Duan brings a stellar track record from his PhD in Robotics and AI at the University of Washington, where he was co-advised by renowned researchers Ranjay Krishna and Dieter Fox. A Nanyang Technological University (NTU) alumnus with a BEng in Electrical and Electronic Engineering (Highest Distinction), Duan's journey reflects Singapore's talent pipeline from local education to global excellence.

Duan's research has garnered accolades, including Best Paper Awards at Ubiquitous Robots 2023 and RememberRL@CoRL 2025, Spotlight at ICLR 2024, and Oral presentations at RSS and ICRA. His work has been highlighted in MIT Technology Review and VentureBeat. Previously a Graduate Student Researcher at the Allen Institute for AI and an intern at NVIDIA, Duan is passionate about robot learning and large-scale robotics foundation models. "Our goal is to build the essential blocks for AI in the physical world, ensuring readiness for real-world deployment," he states on his personal site.

Under Duan's direction, MAGIC Lab emphasizes human-centric robotics, bridging academic theory with practical applications. This leadership positions NUS at the forefront of Singapore's higher education push in AI-driven robotics.

Core Research Areas and Breakthrough Projects

MAGIC Lab's research spans multimodal large language model (MLLM) reasoning, 3D vision, robot learning, simulation environments, and dexterous manipulation. These areas tackle key hurdles in embodied AI, such as enabling robots to generalize across tasks and environments without extensive retraining.

Recent projects exemplify the lab's innovation. MolmoAct2, released in 2026, advances action reasoning models for real-world robotic deployment. MolmoSpace, accepted to RSS 2026, creates large-scale ecosystems for robot navigation and manipulation. Other highlights include SAM2Act for visual foundation models in manipulation (ICML 2025) and Manipulate-Anything (CoRL 2024), which automates robots using vision-language models. Publications in top venues like ICLR, CoRL, RSS, and ICRA underscore the lab's rigorous output.

The lab employs cross-embodiment learning to transfer skills between robot types, accelerating development. For instance, Selective Visual Representations (ICLR 2024 Spotlight) improves convergence in embodied AI tasks. These efforts build on Duan's prior works like AR2-D2 (CoRL 2023), training robots entirely in simulation. Detailed publications are available on the lab's official page.

This research not only pushes theoretical boundaries but also aligns with industry needs, such as agile manufacturing robots capable of handling unstructured environments.

Building a Dynamic Team: Current Members and Open Positions

MAGIC Lab starts with a talented core team, including Haoquan Fang (incoming Stanford PhD), Shuo Liu (UW MS), and others from top programs like UPenn and CMU. Alumni have secured roles at Google, Shopee, MIT, and AI2, highlighting the lab's mentorship strength.

The lab is aggressively recruiting PhD students, postdocs, research assistants (RAs), master's students, and undergraduates. PhD, postdoc, and RA positions offer full funding, making them highly competitive. Robotics experience is preferred but not mandatory; applicants from diverse backgrounds are encouraged. Interested candidates should complete the interest form and follow up via email.

This expansion reflects NUS's strategy to attract global talent, supporting Singapore's goal of 15,000 AI practitioners by 2030.

Photo by Markus Spiske on Unsplash

Integration within NUS's Robotics and AI Ecosystem

MAGIC Lab complements NUS's established centers. The Advanced Robotics Centre (ARC), founded in 2013 by NUS College of Design & Engineering and School of Computing, advances robotics for industry and quality-of-life improvements. ARC focuses on novel technologies with high impact.

NUS AI Singapore Lab, awarded S$9.9 million in 2021, drives embodied, interactive, and trustworthy AI. Collaborations with Microsoft Research Asia and Grab further amplify efforts. NUS Computing's PhD in Computer Science program, emphasizing AI, provides a pipeline for MAGIC Lab recruits.

These synergies position NUS as Singapore's premier hub for robotics higher education.

Singapore's National Landscape: NTU and Government Initiatives

While NUS leads in computing-centric robotics, Nanyang Technological University (NTU) excels in engineering applications. NTU's Robotics Research Centre (RRC), established 1994, and the Schaeffler-NTU Corporate Lab (launched 2025) target AI-enabled humanoid robots for manufacturing and healthcare.

Government backing is robust. The National Robotics Programme (NRP) invests S$635 million, bridging research and market via RoboCluster and RoboNexus. The S$1 billion National AI R&D Plan (2025-2030) funds infrastructure and talent. Initiatives like AI Singapore foster university-industry ties, with NUS and NTU central.

This ecosystem ensures Singapore universities produce graduates ready for a robotics-driven economy.

Educational Programs and Student Opportunities at NUS

NUS School of Computing offers a PhD in Computer Science with AI specializations, where students take advanced courses before dissertation research. The BEng in Robotics & Machine Intelligence equips undergraduates with skills in intelligent systems.

MAGIC Lab provides hands-on opportunities, from RA roles to PhD supervision. Students gain exposure to cutting-edge tools, simulations, and real-world deployments, enhancing employability in Singapore's tech sector, projected to need 20,000 AI roles by 2026.

• PhD funding: Full scholarships covering tuition and stipend.
• Master's projects: Collaborative with industry partners.
• Undergrad internships: Credit-bearing research attachments.

Career Prospects in Singapore's AI Robotics Sector

Graduates from NUS MAGIC Lab enter a booming market. Singapore's robotics industry is growing at 15% annually, driven by automation in ports, hospitals, and factories. Roles include robotics engineers (S$80,000-120,000 starting), AI researchers, and PhD postdocs.

Explore opportunities via platforms like NUS research jobs. The lab's industry-relevant projects prepare students for firms like ST Engineering and A*STAR.

Photo by TSquared Lab on Unsplash

Future Outlook: Impacts and Challenges

MAGIC Lab's work promises advancements in dexterous robots for elderly care and precision manufacturing, aligning with Singapore's aging population and Industry 4.0. Challenges include ethical AI deployment and talent retention amid global competition.

With government support, NUS aims for breakthroughs by 2030, potentially spinning off startups via NUS Enterprise's S$150 million VC program.

In summary, MAGIC Lab elevates NUS and Singapore higher education, fostering innovation that bridges academia and society. Aspiring researchers should seize these opportunities to shape the future of robotics.