OSCAR: Open Systems Challenges — Agenda and Roadmap

Co-located with EuroSys 2026

OSCAR Workshop at EuroSys 2026, Edinburgh
   
Date April 27, 2026
Time 14:00 – 17:30
Room Moorfoot

Abstract

Modern computing systems are evolving rapidly, driven by the growth of large-scale data centres, heterogeneous hardware, and increasingly complex AI and distributed workloads. While systems research continues to produce important advances, many of the most impactful problems emerge at the intersection of research and real-world deployments. Identifying and articulating these problems clearly is essential for guiding future research directions.

OSCAR (Open Systems Challenges — Agenda and Roadmap) is a discussion-focused workshop aimed at identifying and clarifying the most important open challenges in systems research. The workshop brings together leading researchers from academia and industry to present focused talks outlining concrete problems that remain unsolved in practice, and where new systems ideas could have significant impact.

Rather than a traditional paper workshop, OSCAR is designed as an invitation-based forum centred on problem formulation and discussion. Speakers will highlight emerging challenges across areas such as operating systems, distributed systems, data centre infrastructure, AI/ML systems, and heterogeneous computing. The goal is to move beyond incremental improvements and instead articulate research questions that can shape the next decade of systems research.

A central aim of the workshop is to strengthen the dialogue between academic and industrial systems communities. By surfacing real-world challenges and framing them as clear research opportunities, OSCAR seeks to help guide PhD students, early-career researchers, and the broader systems community toward problems that matter for both scientific progress and practical impact.


Program

Time Topic Speaker & Institution
14:00 – 14:45 The Aversion to Systematization in Systems Research Prof. Timothy Roscoe
ETH Zürich
14:45 – 15:05 From Paper to Community and Deployment: Lessons from Building RL Infrastructure Guangming Sheng
University of Hongkong
15:05 – 15:45 AI Systems Research in a Vibe Coding Era Prof. Kenneth P. Birman
Cornell University
15:45 – 16:05 Coffee Break
16:05 – 16:45 Towards A Learning-Directed Operating System Prof. Aditya Akella
University of Texas at Austin
16:45 – 17:30 How Will AI Systems Do Systems Research? Ant Rowstron
CTO at ARIA

Workshop Committee

Prof. Adam Barker

University of Edinburgh

Prof. Boris Grot

University of Edinburgh

Prof. Ahmed Ali-Eldin Hassan

Chalmers University of Technology