I build systems that make AI faster, more efficient, and easier to run. My research brings together LLM serving, multi-agent infrastructure, and algorithms that learn from the models they support.
I am a postdoctoral fellow at MIT CSAIL. Previously, I was a postdoctoral fellow at Harvard University. I received my Ph.D. in Computer Science from the Technion, where I also completed my B.Sc. and M.Sc. Before academia, I worked as a senior software engineer at Mellanox, now part of NVIDIA.
News
-
Orla, our library for serving LLM-based multi-agent systems, was accepted to CAIS 2026 and received Slingshot funding from the Laude Institute.
-
FIELDING: Clustered Federated Learning with Data Drift was accepted to AISTATS 2026.
-
MARS: Fast Inference for Augmented Large Language Models was accepted to NeurIPS 2025.
-
We presented a tutorial at SIGMETRICS 2025 on queueing systems, predictions, and LLM inference.