Applied AI Systems & Agents
Applied AI Systems & Agents chapter is dedicated to building the agent infrastructure that will turn IAIC's research and engineering capabilities into AI systems deployable across the six application domains—Semiconductor, Advanced Manufacturing, Sustainability, Hub of the Future, Healthcare, and Digital Services.
Our work spans the full agent lifecycle: specification-driven design (SDD), harness engineering, and pattern-based composition drawing from a library of validated agent design patterns. The deliverables include defect detection and process optimisation for Semiconductor; quality inspection and predictive maintenance for Advanced Manufacturing; carbon tracking and energy optimisation for Decarbonisation; urban planning and transport orchestration for Hub of the Future; clinical summarisation and decision support for Healthcare; and natural-language data query and compliance monitoring for Finance & Digital Services.
Beyond building agents one at a time, we are exploring a deeper transformation: how agentic systems reshape the very pattern of research and engineering work—turning sequential, human-driven processes into parallel, agent-orchestrated workflows where cognition, coordination, and a shared runtime substrate compress weeks of effort into hours. Aligned with A*STAR's AI-native ambition, we operate as a horizontal capability: any of the six application domains can pull our agent layer together with Data, Model, Safety, and Compute chapters to form one cross-functional delivery squad.
Data & Knowledge
Data and institutional knowledge distilled from past work are essential lubricants for turning the innovation flywheel in any scientific organisation. They are first-class assets that will enable IAIC to go from strength to strength in the long road to research excellence.
A research organisation that can unlock the latent value in its data and institutional knowledge via a virtuous cycle of accumulation and reuse is one that is poised to iterate faster and explore broader, compounding the rewards of past work into future opportunities.
The Data & Knowledge chapter focuses on making data reusable and knowledge accessible to all in IAIC with proper data architecture and governance. Its purpose is to provide IAIC’s research talents with fuel to innovate fearlessly and stand at the forefront of research by providing not only policy guidance in the proper use (and reuse) of scientific data but also the technology infrastructure that would make data access easy and seamless.
Embodied AI
The Embodied AI & Robotics chapter advances robotics towards systems that can reason, learn from large-scale data, and generalise across diverse real-world environments through embodied experience. This foundation is built by integrating physical self-awareness, dexterous manipulation, and commonsense reasoning into Robot Foundation Models, supported by robust multimodal perception capabilities, particularly tactile and force sensing. Together, these capabilities enable skilled tool use, fast and safe actions, and natural human–robot interaction. A key part of this effort is a data engine that supports the emergence of new skills through compositional learning, efficient adaptation from high-quality data, and cross-task generalisation.
The chapter’s goal is to translate state-of-the-art research into trusted, scalable, multi-functional robotic systems that deliver meaningful real-world impact. As such, safety and trustworthiness are first-class considerations, through compliant control, contact-aware interaction, and embodied reasoning. These are developed and validated through pilots and deployments in high-value sectors such as manufacturing, aviation, and maritime, which are sectors where reliability and adaptability are critical.
High Performance Computing
The High Performance Computing chapter provides the shared compute and systems capabilities that underpin IAIC’s Data-AI-Compute (DAC) mission across its research and application pillars. As a horizontal capability layer, the chapter works closely with domain teams and partner chapters to advance foundational and applied research in compute and systems, and to translate advances in data, AI, and modelling and simulation (M&S) into scalable, efficient, and deployable solutions. Its focus is on building reusable platforms, shared methods, and execution capabilities that can be adapted across multiple application areas, helping domain teams develop and deliver end-to-end solutions.
It spans classical and emerging computing paradigms, supporting workloads from AI and M&S to digital twins and future hybrid quantum-classical workflows. Through this role, the Compute & Systems chapter helps IAIC accelerate research, improve compute efficiency, shorten time-to-solution, and enable first-time-right and same-day intelligence, while creating greater value from Data, AI, and Compute investments.
Model & Intelligence
The Model & Intelligence chapter focuses on advancing the development of cutting‑edge AI models and accelerating AI innovation by building scalable, reusable, and efficient model capabilities to support multiple research and application pillars. The Chapter aims to streamline the end‑to‑end AI model development lifecycle, encompassing pre‑training, post‑training, fine‑tuning, and deployment, to ensure consistency, efficiency, and robustness across different domains and projects.
By strengthening and systematising model development practices, we will empower teams to rapidly prototype, adapt, and deploy AI models, enabling faster translation of research into real‑world impact in dynamic and evolving environments. Building on this foundation, the Chapter serves as a core enabling function across IAIC, driving model excellence, accelerating the transition from research to deployment, and supporting sustainable AI innovation at scale.
Safety & Security
The Security and Safety chapter advances applied research and development to build trustworthy, secure, and resilient digital systems. We focus on translating deep-tech innovations into practical solutions that address real-world risks across enterprises, critical infrastructure, and emerging autonomous ecosystems.
Our work spans the full R&D lifecycle—from early-stage exploration and prototyping to system integration, validation, and deployment. A core focus is trustworthy AI, including AI safety, governance frameworks, and rigorous red-teaming of AI-enabled systems to identify and mitigate vulnerabilities before deployment.
In parallel, we develop and apply AI-driven cybersecurity capabilities to detect, prevent, and respond to evolving threats across enterprise environments, next-generation mobile networks (5G and beyond, including 6G), and operational technology (OT) systems. We also leverage agentic AI technologies to enhance Security Operations Centres (SOC) and Network Operations Centres (NOC), enabling more autonomous, scalable, and intelligent threat monitoring and response. Our research further addresses cybersecurity and privacy in autonomous systems, such as UAVs, USVs, and humanoid robots, and includes technologies to combat scams, misinformation, and online fraud.
Beyond research, we actively drive translation through industry partnerships, with pathways for technology licensing, spin-offs, and startup creation to bring impactful solutions to market.