Workshop on ML for Systems at NeurIPS 2026, December 11 or 12 (TBA), International Convention Center Sydney
Workshop on ML for Systems at NeurIPS '26, Sydney, Australia

Workshop Overview

Machine Learning for Systems (ML for Systems) is an annual, interdisciplinary workshop that brings together researchers and practitioners in computer systems and machine learning. The workshop focuses on applying machine learning techniques to the design, optimization, and operation of computer systems, while also exploring how emerging ML paradigms—including large language models (LLMs), multimodal foundation models, and agentic workflows—can be leveraged to address systems challenges and improve the efficiency, reliability, and scalability of ML infrastructure itself.

Important Dates

  • Paper Submission Deadline: August 29, 2026 by midnight (Anywhere on Earth).
  • Paper Acceptance Notifications: September 29, 2026
  • Workshop: December 11 or 12, 2026 (To be announced)

Call for Papers

We invite submissions of up to 4-page extended abstracts describing novel research, systems, datasets, simulators, benchmarks, and methodologies in the broad area of ML for Systems. We are particularly interested in work that advances the state of the art beyond replacing hand-crafted heuristics with learned models, and instead develops principled approaches for designing, optimizing, operating, and evaluating complex systems.

This year, we particularly encourage submissions on:

  • Using LLMs and agentic workflows to address complex systems problems, including program synthesis for hardware design, compiler autotuning, adaptive runtime optimization, system debugging, design-space exploration, and other domains where systems objectives must be balanced with constraints such as reliability, correctness, consistency, and efficiency.
  • Using ML to address emerging systems challenges in large-scale AI infrastructure, including training and serving of LLMs, multimodal models, and agentic applications. Relevant topics include compiler partitioning strategies for distributed training, memory and compute allocation, workflow orchestration, resource scheduling, inference optimization, automated infrastructure management, and efficient utilization of heterogeneous accelerators.
  • Developing best practices, methodologies, benchmarks, datasets, simulators, and evaluation frameworks that improve rigor, reproducibility, reliability, and trustworthiness in ML for Systems research.

Accepted papers will be optionally linked on the workshop website, but there will be no formal proceedings. Authors may therefore publish their work in other journals or conferences. The workshop will include invited talks from industry and academia, as well as oral and poster presentations by workshop participants.

This will be the 10th edition of the workshop. You can find accepted papers from the previous iterations of ML for Systems from NeurIPS 2025, NeurIPS 2024, NeurIPS 2023, NeurIPS 2022, 2021, 2020, 2019, 2018, and ISCA 2019.

Areas of interest:

ML for Systems focuses on the novel application of machine learning techniques towards computer systems problems, for example:

  • Supervised, unsupervised, reinforcement learning, and generative AI (e.g., LLMs) research with applications to:
    • Systems software
    • Runtime systems
    • Distributed systems
    • Security
    • Compilers, data structures, and code optimization
    • Databases
    • Computer architecture, microarchitecture, and accelerators
    • Circuit design and layout
    • Interconnects and Networking
    • Storage
    • Data Centers
    • Programming Languages
  • Representation learning for hardware and software
  • Optimization of computer systems and software
  • Benchmarking and evaluation methodologies for ML for Systems
  • Datasets and simulators for systems optimization and management
  • Reliability, robustness, safety, and trustworthiness of ML-driven systems
  • LLMs and agentic workflows for systems design and optimization
  • Systems for LLMs, multimodal models, and agentic applications
  • Sustainable and energy-efficient computing
  • Emerging applications
    • Using LLMs and agentic workflows for systems
    • Using ML for challenges in large-scale machine learning systems, e.g., training, post-training and serving of LLMs, reasoning models, and agentic workflows
    • Reliability and security challenges in LLM-based agents in systems

Submission Instructions

We welcome submissions of up to 4 pages, not including references or Appendices. This year, this is a strict limit. Authors are welcome to put additional material in the optional Appendix section, but reviewers are not required to read the Appendix.

All submissions must be in PDF format and should follow the NeurIPS 2026 format.

Contact Us

Contact us at mlforsystems@googlegroups.com.