Research
IIDAI researchers advance the state-of-the-art across the following areas: AI Systems, AI-Driven Hardware-Software-System Codesign, Quantum-Centric Supercomputing (QCSC), and New Algorithms Discovery. IIDAI’s body of work addresses challenges in efficient, scalable, and adaptable distributed inference systems to support next-generation AI workloads across diverse computing infrastructures, accelerating the integration of algorithms into silicon for specialized systems, discovering new algorithms that enable classical and quantum systems to work together on problems neither can solve alone, and improving quantum hardware. Our goal is to push the frontiers of AI and quantum-centric computing.
AI Systems
We focus on innovations for advanced AI systems encompassing distributed inference, multi-agent system design and optimization, intelligent memory and KV-cache management, efficient data movement, SLO-driven workload adaptation, integration of reinforcement-learning-based training and inferencing, novel model architectures, and hardware-aware compiler optimization, while exploring systems that increasingly use AI to optimize and evolve themselves.
Algorithms-to-Silicon-to-Systems (AS2)
AS2 captures our AI-Driven Hardware-Software-System Codesign research, promoting an AI-native paradigm that unifies algorithm design, hardware realization, software enablement, and system deployment. AI and agentic systems automate RTL hardware design generation, optimization, verification, compilers, kernels, and system integration, enabling hardware and software to co-evolve. AS2 promises major gains in productivity, scalability, correctness, robustness, and real-world usability.
QCSC Software
We explore the software foundation for tightly integrating quantum and classical high-performance computing. We focus on workflow tools, quantum-aware job scheduling, resource and execution optimization, programming languages and intermediate representations, HPC libraries for quantum-HPC applications, and new coupling models that enable efficient orchestration of quantum and classical computing resources.
QCSC Algorithms & Applications and New Algorithms Discovery
QCSC Algorithms & Applications and New Algorithms Discovery: Our goal is to devise new quantum and quantum-classical algorithms and demonstrate their value on challenging scientific and industrial problems. We explore early fault-tolerant quantum computing algorithms, identify promising new algorithmic approaches, and develop compelling applications and industry use cases, including areas such as drug discovery and new materials design, where quantum-centric supercomputing can provide new capabilities.
Quantum Hardware
Our work on quantum hardware advances the performance, reliability, and scalability of quantum computing systems by addressing fundamental hardware challenges, including noise, loss, and qubit coherence. Our research explores new device structures, circuits, and mitigation techniques to improve quantum processor quality and enable increasingly capable quantum systems.