Zishen Wan
Columbia University
The tutorial
AI agents as
computer architects.
How can AI agents help us design, optimize, and evaluate computing systems? Architecture 2.0 brings this question to IISWC through a half-day program led by our organizing team.
We will introduce QuArch and Archipedia, explore AI-native system design, and present ArchEval and ArchOrchestra. The tutorial connects architectural knowledge, workload analysis, and agent evaluation with the tools needed to study AI-assisted design.
A keynote will offer an invited perspective on the field. Researchers, students, and practitioners interested in agentic AI and computer systems are welcome.
WhenSunday, September 278:00 AM–12:30 PM · Mountain Daylight Time
WhereThe Mall · Embassy Suites Boulder2601 Canyon Blvd, Boulder, CO
Sunday, September 27
Schedule
All times are local to Boulder (MDT, UTC−6). The schedule is subject to change.
| Time (MDT) | Session | Presenter |
|---|---|---|
| 08:00–08:15 | Architecture 2.0 Opening | Zishen Wan |
| 08:15–09:00 | AI-Native System Design | Vijay Janapa Reddi |
| 09:00–10:00 | Keynote TBD | TBD |
| 10:00–10:30 | Coffee Break | — |
| 10:30–11:00 | QuArch & Archipedia: A Benchmark for Evaluating LLM Reasoning in Computer Architecture | Shvetank Prakash |
| 11:00–11:30 | ArchEval: A Composable Measurement Platform for AI Agents as Computer Architects | Chenyu Wang |
| 11:30–12:00 | ArchOrchestra: Orchestrating Accelerator Design via Agent-Driven Exploration | Zishen Wan |
| 12:00–12:30 | Agentic AI for Kernel Synthesis | Ankita Nayak |
Keynote speaker and talk title to be announced.
Explore Architecture 2.0
Resources
Books & Reading Guides
Papers & Articles
- Architecture 2.0: Foundations of Artificial Intelligence Agents for Modern Computer System Design
- GenAI for Systems: Recurring Challenges and Design Principles from Software to Silicon
- ArchAgent: Agentic AI-driven Computer Architecture Discovery
- Computer Architecture’s AlphaZero Moment: Automated Discovery in an Encircled World
- Agentic Architect: An Agentic AI Framework for Architecture Design Exploration and Optimization
- CHIA: An Open-Source Framework for Principled, Agentic AI-Driven Hardware/Software Co-Design Research
Perspectives & Blog Posts
Benchmarks & Datasets
Tools & Design Infrastructure
Courses & Community
Organizers & tutorial team
Organizing team
Chenyu Wang
Harvard University
Shvetank Prakash
Harvard University
Andy Cheng
Harvard University
Arya Tschand
Harvard University
Ankita Nayak
Gimlet Labs
Vijay Janapa Reddi
Harvard University