Zishen Wan

Assistant Professor of Computer Science · Columbia University

I am a computer architect and SoC designer working across computer architecture, systems, and chip/VLSI design. My group develops cross-layer computing systems that span software system, hardware architecture, and silicon for emerging AI workloads, including embodied, physical, neuro-symbolic, and agentic AI. I also develop agentic AI methods for the design, optimization, and verification of computing systems.

Before joining Columbia University, I was a Postdoctoral Fellow at Harvard University, working with Prof. Vijay Janapa Reddi. I received my Ph.D. from Georgia Tech in 2025, advised by Prof. Arijit Raychowdhury and Prof. Tushar Krishna.

Selected honors. My research has been recognized with Best Paper Awards from DAC, IEEE Computer Architecture Letters (CAL), and SRC JUMP 2.0, as well as IEEE Micro Top Picks and ACM SIGDA Research Highlights. My Ph.D. dissertation received ACM SIGDA Outstanding Ph.D. Dissertation Award, ACM FCCM Outstanding Ph.D. Dissertation Award, and Georgia Tech's Colonel Oscar P. Cleaver Award. I was also awarded first place at DAC Ph.D. Forum and ACM Student Research Competition, WAIC Yunfan Award, and selected as both ML and Systems Rising Star and Cyber-Physical Systems Rising Star.

Prospective students and postdocs: I am recruiting Ph.D. students, postdoctoral researchers, and visiting students interested in computer architecture, systems, and chip/VLSI design. Get in touch →

Research

My research studies how to build computing systems for emerging intelligence and how AI can help design computing systems. These two complementary directions span computer architecture, computer systems, and chip/VLSI design. Across both directions, I use cross-layer co-design to improve efficiency, scalability, reliability, and adaptability.

Research Direction

Computing for AI

I design systems, architectures, and chips for emerging workloads, including embodied, physical, neuro-symbolic, and agentic AI.

Research Direction

AI for Computing

I develop agentic AI methods for design exploration, discovery, optimization, and verification across the computing stack.

Research overview showing Computing for AI and AI for Computing across computer architecture, systems, and chips and VLSI.

Swipe horizontally to explore the full research map

Computer Architecture

Architectures for Emerging Intelligence

Domain-specific and adaptive architectures for embodied, neuro-symbolic, and reasoning workloads, including accelerators, memory systems, and heterogeneous platforms.

AI for Computer Architecture: Agentic design space exploration, generation, and evaluation.

Accelerators · Software-hardware co-design · Heterogeneous architecture · Memory · Dataflow

Systems

AI Systems & Cross-Layer Co-Design

System support for emerging intelligence, including workload characterization, runtime and serving systems, hardware and software co-design, and resilience.

AI for Systems: Agentic methods for system modeling, optimization, and operation.

Agentic serving · Workload characterization · Reliability · System co-design · GPU/CPU/NPU/TPU

Chips & VLSI

AI Chips & Memory-Centric Computing

AI SoCs, FPGA prototypes, memory-centric architectures, and emerging-device circuits that translate system and architecture ideas into working silicon.

AI for Chip Design: Agentic methods for hardware generation, optimization, and verification.

SoCs · VLSI Design · FPGA · Non-volatile memory · Memory-centric computing · Emerging devices

Representative Work

View all publications →
Computer Architecture

Architectures for Neuro-Symbolic and Cognitive AI

Summary: Efficient Neuro-Symbolic AI Computing

Architectures for Physical AI and Autonomous Systems

  • ArboristMotion-planning acceleration · MICRO 2026
  • RTGSReal-time 3D Gaussian splatting SLAM · MICRO 2025
  • OctoCache3D occupancy mapping acceleration · ASPLOS 2025
  • Thinking and MovingCooperative embodied-system computing · ICCAD 2024
  • ORIANNARobotic accelerator generation · ASPLOS 2024
  • AutoPilotDomain-specific SoC design for aerial robots · MICRO 2022
  • The Sky Is Not the LimitCyber-physical performance modeling · CAL 2020
Systems

Systems for Physical AI

  • CREATECross-layer resilience for embodied AI · ASPLOS 2026
  • ReCACooperative embodied-agent acceleration · ASPLOS 2025
  • EmbodiedPerfGenerative AI in embodied systems · ISPASS 2025
  • RobotPerfRobotics computing benchmark suite · ICRA 2024

Systems for Agentic AI and Serving

  • DyserveWorkflow-aware serving for agentic applications · SCALE at ICML 2026
  • Faster-MoALow-latency mixture-of-agents serving · DAC 2026
  • SLM-MuxSmall language model orchestration for reasoning · ICLR 2026

Workload Characterization and System Resilience

  • Towards System-2 AIEnergy-based model characterization · ISPASS 2026
  • FortiSkyReliable autonomous-machine systems · DATE 2026
  • ReaLMReliable and efficient LLM inference · DAC 2025
  • MulBERRYBit-error robustness for multi-agent systems · ASPLOS 2024
  • VAPProtection for reliable autonomous machines · CACM 2024
  • MAVFIEnd-to-end fault analysis for aerial robots · DATE and TCAD 2023
Chips & VLSI

AI SoCs and Tapeouts

Memory-Centric and Emerging-Device Computing

Agentic AI for Computing System Design
  • ArchEvalMeasuring AI agents as computer architects · ISCA Workshop 2026
  • AgentDSEReasoning-augmented architectural exploration · ISCA Workshop 2026
  • GenAI for SystemsDesign principles from software to silicon · arXiv 2026
  • LACELLM-aided RISC-V instruction extension · ICLAD 2026
  • QuArchEvaluating LLM reasoning in computer architecture · ICML 2026
  • HDLxGraphLLM access to HDL repositories through graph databases · ASP-DAC 2026

News

News Archive

Books

Cover of Robotic Computing on FPGAs
Co-author

Robotic Computing on FPGAs

Synthesis Lectures on Computer Architecture · Springer · 2021

An overview of FPGA accelerator design for robotic perception, localization, planning, and multi-robot collaboration. The book connects architectural techniques with deployments in autonomous vehicles and space robotics.

Book
Cover of Machine Learning Systems
Contributing Author

Machine Learning Systems: Principles and Practices of Engineering Artificially Intelligent Systems

Open-source textbook · Harvard Edge Computing Lab · 2025 to present

An open-source textbook on engineering end-to-end machine learning systems, from data and modeling to deployment, acceleration, security, reliability, and responsible AI. Developed by the Harvard Edge Computing Lab community and used in the CS249r course.

Book GitHub Contributors

Ph.D. Dissertation

Cover of Tailored Computing doctoral dissertation
Doctoral Dissertation

Tailored Computing: Cross-Layer System, Architecture, and Silicon Co-Design for Physical AI

Zishen Wan · Georgia Institute of Technology · 2025

This dissertation develops a cross-layer system, architecture, and silicon co-design framework for efficient, reliable, and scalable physical intelligence. It connects neuro-symbolic reasoning and embodied AI with unified abstractions, domain-specific architectures, and working chip implementations.

Selected Recognition

  • ACM SIGDA Outstanding Ph.D. Dissertation Award, 2026
  • Georgia Tech Colonel Oscar Cleaver Dissertation Award, 2026
  • ACM FCCM Outstanding Ph.D. Dissertation Award, 2026
  • First Place, ACM/IEEE DAC PhD Forum, 2025
  • First Place, ACM Student Research Competition, 2024

Publications

A curated collection of representative papers across computer architecture, systems, chips and VLSI, and AI for computing-system design.

Google Scholar
Topics:

GenAI for Systems: Recurring Challenges and Design Principles from Software to Silicon

Arya Tschand*, Chenyu Wang*, Zishen Wan*, Andrew Cheng, Ioana Cristescu, Kevin He, Howard Huang, Alexander Ingare, Akseli Kangaslahti, Sara Kangaslahti, Theo Lebryk, Hongjin Lin, Jeffrey Jian Ma, Alexandru Meterez, Clara Mohri, Depen Morwani, Sunny Qin, Roy Rinberg, Paula Rodriguez-Diaz, Alyssa Mia Taliotis, Pernille Undrum Fathi, Rosie Zhao, Todd Zhou, Vijay Janapa Reddi

2026

arXiv 2026

PaperWebsite

QuArch: A Benchmark for Evaluating LLM Reasoning in Computer Architecture

Shvetank Prakash, Andy Cheng, Arya Tschand, Mark Mazumder, Varun Gohil, Jeff Ma, Jason Yik, Zishen Wan, Jessica Quaye, Elisavet Lydia Alvanaki, Avinash Kumar, Chandrashis Mazumdar, Tuhin Khare, Alexander Ingare, Ikechukwu Uchendu, Radhika Ghosal, Abhishek Tyagi, Chenyu Wang, Andrea Mattia Garavagno, Sarah Gu, Alice Guo, Grace Hur, Luca Carloni, Tushar Krishna, Ankita Nayak, Amir Yazdanbakhsh, Vijay Janapa Reddi

2026

ICML 2026

PaperProject Website

RobotPerf: An Open-Source, Vendor-Agnostic, Benchmarking Suite for Evaluating Robotics Computing System Performance

Victor Mayoral-Vilches, Jason Jabbour, Yu-Shun Hsiao, Zishen Wan, Alejandra Martinez-Farina, Martino Crespo-Alvarez, Matthew Stewart, Juan Manuel Reina-Munoz, Prateek Nagras, Gaurav Vikhe, Mohammad Bakhshalipour, Martin Pinzger, Stefan Rass, Smruti Panigrahi, Giulio Corradi, Niladri Roy, Phillip B. Gibbons, Sabrina M. Neuman, Brian Plancher, Vijay Janapa Reddi

2024

ICRA 2024Best Paper Award, IROS Robotics Benchmarking Workshop

PaperPosterCodeProject PageMedia

Honors and Awards

Selected distinctions in research, scholarship, and academic leadership.

  1. ACM SIGDA Outstanding Ph.D. Dissertation Award
  2. ACM FCCM Outstanding Ph.D. Dissertation Award
  3. Colonel Oscar P. Cleaver Ph.D. Dissertation Award
  4. WAIC Yunfan Award
  5. ACM Future AI LeaderACM AI Leadership Summit
  6. First Place, DAC Ph.D. Forum
  7. Best Paper AwardDARPA SRC JUMP 2.0 Program
  8. Baidu Ph.D. Fellowship
  9. Cyber-Physical Systems Rising Star
  10. Best Paper AwardSemiconductor Research Corporation TECHCON
  11. Machine Learning and Systems Rising Star
  12. IEEE Micro Top Picks
  13. First Place, ACM Student Research Competition
  14. Best Paper AwardIROS Robotics Benchmarking Workshop
  15. Qualcomm Fellowship
  16. CRNCH Ph.D. Fellowship
  17. Best Paper AwardIEEE/ACM Design Automation Conference
  18. Best Paper AwardIEEE Computer Architecture Letters
Additional honors
  1. Best Poster AwardDARPA SRC JUMP 2.0 CoCoSys Annual Review
  2. Best Paper AwardDARPA SRC JUMP 2.0 Program
  3. Best Poster AwardDARPA SRC JUMP 2.0 CoCoSys Annual Review
  4. Third Place, ACM SIGMICRO Student Research CompetitionIEEE/ACM International Symposium on Microarchitecture
  5. CoCoSys Spotlight Research Scholar
  6. Best Poster AwardIBM IEEE AI Compute Symposium
  7. Roger P. Webb Graduate Research Assistant Excellence AwardGeorgia Tech
  8. DAC Young Fellow
  9. ACM SIGDA Research Highlights Nominee
  10. DAC Young Fellow
  11. Best Presentation AwardDAC Young Fellow Program

Selected Talks

Invited seminars and research forums, organized by talk series.

AI-Native Computing Systems: From Computing for AI to AI for Computing 2 talks
  • Seminar Talk, Harvard UniversityHosted by Prof. Juncheng Yang and Prof. Minlan Yu
  • Berkeley Agentic AI SummitUniversity of California, Berkeley
Architecting Physical Intelligence: Cross-Stack Co-Design from Systems to Silicon 23 institutions

Arizona State University, Boston University, Chinese University of Hong Kong, Columbia University, Harvard University, Hong Kong University of Science and Technology, Johns Hopkins University, Mohamed bin Zayed University of Artificial Intelligence, National University of Singapore, North Carolina State University, Northeastern University, Peking University, Purdue University, Rice University, Texas A&M University, Tsinghua University, University of Colorado Boulder, University of Illinois Urbana-Champaign, University of Maryland, University of Pennsylvania, University of Southern California, and University of Texas.

Tailored Computing: Cross-Layer System, Architecture, and Silicon Co-Design for Embodied Autonomous Machines 8 talks
  • Tsinghua University NICS-EFC LabInvited talk hosted by Dr. Zhenhua Zhu
  • MICRO Ph.D. ForumSeoul, Korea
  • DAC Ph.D. ForumSan Francisco, California
  • ISCA Workshop on Architecture Support for Embodied AI Systems
  • CoCoSys Annual SummitDARPA SRC JUMP 2.0, Atlanta, Georgia
  • UIUC Coordinated Science LaboratoryInvited talk, Champaign, Illinois
  • University of WashingtonSeminar hosted by Prof. Ang Li
  • Institute of Computing Technology, Chinese Academy of SciencesSeminar hosted by Prof. Yunji Chen
System Implications and Opportunities for Compositional Neuro-Symbolic-Probabilistic AI 1 talk
  • Georgia TechSeminar hosted by Prof. Alexey Tumanov, Atlanta, Georgia
Demystifying Neuro-Symbolic AI via Workload Characterization and Software-Hardware Co-Design 14 talks
  • Purdue UniversitySeminar hosted by Prof. Anand Raghunathan and Prof. Kaushik Roy
  • University of Notre DameSeminar hosted by Prof. Ningyuan Cao
  • GoogleHosted by Dr. Suvinay Subramanian
  • CoCoSys Annual SummitDARPA SRC JUMP 2.0
  • Georgia Tech ECE8893Guest lecture hosted by Prof. Callie Hao
  • Georgia Tech Computer Architecture Research Seminar
  • ACM Student Research Competition at MICROAustin, Texas
  • University of Minnesota Twin CitiesInvited talk hosted by Prof. Katie Zhao
  • MLSys Young Professional SymposiumSanta Clara, California
  • International Workshop on Neuro-Symbolic SystemsUniversity of California, Berkeley
  • CoCoSys Annual SummitDARPA SRC JUMP 2.0
  • Kennesaw State University EE6900Guest lecture hosted by Prof. Yan Fang
  • Georgia Tech 3D Systems Packaging Research Center
  • CoCoSys Annual SummitDARPA SRC JUMP 2.0
System-Architecture-Technology Cross-Layer Design for Embodied Intelligence 2 talks
  • Harvard University Nano-Design Research GroupHosted by Prof. Gage Hills
  • Lawrence Livermore National LaboratoryHosted by Dr. Kshitij Bhardwaj
Intelligence in Robotic Computing: Agile Design Flows for Efficient Autonomous Systems 13 talks
  • University of Central Florida Computer Architecture SeminarHosted by Prof. Di Wu
  • MICRO Workshop on Robotics Acceleration with Computing Hardware
  • ESWEEK Ph.D. ForumRaleigh, North Carolina
  • Cyber-Physical Systems Rising Stars WorkshopUniversity of Virginia
  • CoCoSys Liaison MeetingDARPA SRC JUMP 2.0
  • Georgia Tech CRIDC
  • Georgia Tech Computer Architecture Research Seminar
  • IBM AI Compute SymposiumIBM T.J. Watson Research Center
  • ML and Systems Rising Stars WorkshopGoogle, Mountain View, California
  • Georgia Tech Chips Day
  • Georgia Tech EIC LabHosted by Prof. Celine Lin
  • CRNCH Annual SummitGeorgia Tech
  • ACM Student Research Competition at ICCADSan Diego, California
Efficient Algorithm-Hardware Co-Design for Robotic Mapping and Localization 6 talks
  • Georgia Tech Computer Architecture Research Seminar
  • Georgia Tech CRIDC
  • IBM AI Compute SymposiumIBM T.J. Watson Research Center
  • CBRIC Annual SummitDARPA JUMP SRC, Purdue University
  • Georgia Tech ECE8893Guest lecture hosted by Prof. Callie Hao
  • CRNCH Annual Summit
Enabling Reliable and Safe Autonomous Systems 8 talks
  • CoCoSys Annual SummitDARPA SRC JUMP 2.0
  • CRNCH Annual Summit
  • CoCoSys Annual SummitDARPA SRC JUMP 2.0
  • ACM Student Research Competition at ESWEEK
  • IEEE COMPSAC Plenary PanelReliability of Autonomous Machines
  • CBRIC Annual SummitDARPA JUMP SRC
  • CBRIC Industry TalkDARPA JUMP SRC
  • Harvard Architecture, Circuits and Compilers Lab

Teaching

Courses and lectures across computer architecture, AI systems, and hardware design.

Course Development and Teaching Staff

CS249r: Architecture 2.0

Agentic AI for Computing Systems Design

Harvard University Fall 2025
Guest Lectures
  • Software-Hardware Co-Design for Neuro-Symbolic AIGeorgia Tech ECE8893, Parallel Programming for FPGAs
  • Opportunities and Challenges for In-Memory and Near-Memory ComputingGeorgia Tech ECE8803, Software-Hardware Co-Design for Machine Learning Systems
  • FPGA-Based Robotic ComputingGeorgia Tech ECE8893, Parallel Programming for FPGAs
  • FPGA-Based Robotic ComputingGeorgia Tech ECE8893, Parallel Programming for FPGAs
  • Neuro-Symbolic Computing SystemsKennesaw State University EE6900, Neuromorphic Computing

Academic Service

Leadership and professional service across computer architecture, systems, and design automation.

Conference and Workshop Leadership
  • Organizer, Architecture 2.0: Agentic AI for Computing Systems Design at ASPLOS, ISCA, and IISWC
  • Organizer, VisArch: Visualizations for Computer Architecture at ISCA
  • Special-session organizer, Domain-Specific Hardware for Emerging AI Workloads at ICCAD
  • Special-session organizer, Neuro-Symbolic Architecture Meets Large Language Models at ESWEEK
Technical Program Committees
  • HPCA, ASPLOS, MICRO, DAC, MLSys, ASAP, ICCD, ISLPED, MLCAD
  • MLSys, ICRA
  • DAC
  • DAC, ESWEEK, IJCAI
Artifact Evaluation Committees

HPCA 2026, MLSys 2026, HPCA 2025, ISCA 2024, MICRO 2023, ISCA 2023, ASPLOS 2023, MLSys 2023, MICRO 2022, ASPLOS 2022, IISWC 2022

Workshop Program Committees

MLBench at ASPLOS 2026, Arch4EAI at ISCA 2025, SCOPE at ICLR 2025, Lock-LLM at NeurIPS 2025, CAV at ASPLOS 2024

Journal Reviewing

IEEE JSSC, IEEE TCAD, IEEE TCAS-I, IEEE TCAS-II, IEEE TBioCAS, IEEE JETCAS, IEEE Micro, IEEE Internet of Things Journal, IEEE CAL, IEEE TIM, ACM JATS, ACM TCPS

Community Leadership and Outreach
  • Steering Committee, Computer Architecture Student Association
  • Panelist, ML and Systems Rising Stars Workshop
  • News and Media Team, IEEE ISSCC
  • Co-founder, MLCommons Resilience and Robustness Research Working Group
  • Plenary panelist, IEEE COMPSAC

Media Coverage

Selected coverage of research in AI systems, computer architecture, and autonomous computing.

Institutional and program news