Max Shad
Senior Director, AI/ML Research Engineering
I lead AI/ML research engineering and AI infrastructure & compute at the Kempner Institute at Harvard. My AI/ML team builds multimodal foundation models from scratch, along with the platforms, workflows, complex agentic solutions, and tools that researchers utilize on our AI cluster.
At a glance
- AI Cluster Design
- +1,100 GPUs
- TOP500, US Academic
- 3rd
- Scalable Data Solutions
- +5 PB
- Computing Handbook
- 110+ countries
- Joined Harvard
- 2017
News
All newsModerated the panel Generative & Agentic AI in Research at the RC Summit, Harvard University.
Moderated the panel Challenges in AI Facilitation at the RC Summit, Harvard University.
Blog
All postsWe Made Research Faster. Why Are We Surprised It's Moving Faster?
We need to redesign peer review for abundance.
Comparing Nvidia Data Center GPUs
Measured compute, memory bandwidth, and training throughput on four Nvidia GPUs, and what the differences mean in practice.
InteractiveTechnical
Selected work
All workPublications
All publications
Research Data that Can Be Trusted
M. Bouzinier, D. Etin, N. Khoshnevis, M. Shad, S. Yockel
SpringerBriefs in Computer Science, Springer Cham
Fast Recovery Is Not Enough: Correct Goodput and Trajectory Fidelity for Foundation-Model Training on Preemptible Clusters
N. Khoshnevis, Y. Mazloumi, C. Brown-Pinilla, T. Ngotiaoco, D. Balamurugan, S. Leinicke, M. Shad
IEEE High Performance Extreme Computing Conference (HPEC)
AI/MLHPC