mlsys.org

What It Is

MLSys 2026, the 9th Annual Conference on Machine Learning and Systems, is a leading academic conference focused on the intersection of machine learning and systems. The 2026 conference will be held in Bellevue, Washington, USA, and aims to explore design principles for learning systems, novel learning methods, and theories for practical ML workflows. Its topics cover cutting-edge areas such as large language model (LLM) training and inference, compound AI systems, ML compilers, and specialized hardware.

Key Highlights by Dimension

• Course areas: Spans ML and systems, from low-level hardware acceleration and ML compilers to higher-level LLM fine-tuning, AI Agent systems, data preparation, security, and privacy, covering the full AI infrastructure stack.
• Format: Primarily an in-person conference, including keynotes from experts at top organizations such as NVIDIA and Meta, paper presentations, poster sessions, and competitions. Recordings from past conferences are made freely available about 30 days after the event, providing a learning option for those unable to attend.
• Faculty/institutional background: A strong lineup. The general chair is Luis Ceze from the University of Washington and NVIDIA, while program chairs include Zhihao Jia from CMU and LithosAI. The competition track is sponsored by three major players: AWS, Google, and NVIDIA.
• Certification/certificates: No traditional course certificate is provided, but proof of attendance is available. Accepted papers and competition awards carry authoritative recognition in both academia and industry.

Pricing and Access

The page does not directly disclose conference registration fees, but it mentions that credit card payments are accepted and that discounts are available for full-time students. Additional services, such as on-site poster printing, cost between $65 and $140. For users in China, the accessibility status of the official website is unknown. However, attending in person in the U.S. requires arranging a visa in advance, and the conference only issues invitation letters to registered in-person attendees.

Pros and Cons

• Pros: Excellent integration of academia and industry. In 2026, a dedicated industry paper track has been added to encourage real-world system retrospectives. The agenda is highly forward-looking, directly addressing compute and system optimization challenges in the LLM era. Past conference videos are freely available, benefiting the broader community.
• Cons: As an academic conference rather than a structured course, it has a steep learning curve and requires a strong technical background. Core registration fee information is missing. The barrier to in-person participation is relatively high due to visa and travel requirements.

Who It’s For

Best suited for ML systems researchers, AI infrastructure engineers, large-model training and inference optimization specialists, and master’s or PhD students in related fields. If you are building or optimizing large-scale AI systems, this is a must-follow frontier venue for the industry.