Conferences / MLSys (Conference on Machine Learning and Systems)
MLSys 2026
May 18–22, 2026 · Bellevue, WA, United States
- Conference
- MLSys (Conference on Machine Learning and Systems) Academic
- Dates
- May 18, 2026 – May 22, 2026
- Venue
- —
- Place
- Bellevue, WA, United States
- Format
- In person
- Status
- Past
- Attendance
- —
- Website
- mlsys.org
- Recorded
- 32 sponsors · 9 speaker appearances
- Notes
- speakers: keynotes only (invited talks from the MLSys virtual site). Young Professionals Symposium May 18, conference sessions May 19-21, Industry Day May 22. Venue name not stated on the pages consulted. Session recordings were lost in transit after the conference (mlsys.org notice). Registration prices were served by an interactive calculator (mlsys.org/Conferences/<year>/Pricing) that no longer returns past-year rates and is not captured by the Wayback Machine, so they could not be recovered.
Ticket prices
No ticket prices recorded
Sponsors 32
Title / headline 9
- Amazon Web Services · Diamond
- ByteDance · Diamond
- Capital One · Diamond
- Google · Diamond
- Netpreme · Diamond
- NVIDIA · Diamond
- PyTorch Foundation · Diamond
- Verda · Diamond
- Waymo · Diamond
Speakers and talks 9
| Speaker | Company | Role | Talk | Links |
|---|---|---|---|---|
| Amin Vahdat | Keynote | — | — | |
| Christos Kozyrakis | NVIDIA | Keynote | The Path to Infernece Efficiency | — |
| Esha Choukse | Microsoft | Keynote | Beyond Model Serving: Cross-Stack Co-Design for Agentic Systems | — |
| Lidong Zhou | Microsoft | Keynote | The Next Horizon of Systems: From MLSys to System Intelligence | — |
| Lisa Li | OpenAI | Keynote | Eliciting Language Model Behaviors with Investigator Agents | — |
| Luke Zettlemoyer | University of Washington | Keynote | Rethinking Pretraining: Data and Architecture | — |
| Mark Saroufim | Core Automation | Keynote | When AI Starts Writing Systems Code | — |
| Roger Wang | vLLM | Keynote | Rethinking Open Source Contribution in the Age of AI Agents | — |
| Yuhan Liu | University of Chicago | Keynote | LMCache: An Efficient KV Cache Layer for Enterprise-Scale LLM Inference | — |