Babak Ehteshami Bejnordi
I am a research scientist at Qualcomm AI Research (Senior Staff and Manager). My primary research focus lies in the realm of efficient Deep Learning for Large Language Models (LLMs) and Computer Vision. My recent research works have been in the areas of Efficient LLM deployment, Efficient (Latent) Reasoning, Mixture of Experts, Multi-Task Learning, and Continual Learning. I am leading a team on Conditional Computation for efficient deep learning at Qualcomm AI Research, Amsterdam.
I have been the organizer of the Qualcomm Innovation Fellowship Program in Europe since 2019.
I obtained my PhD at the Diagnostic Image Analysis Group, Radboud University, the Netherlands, where I worked on the development of ML algorithms for breast cancer diagnostics. During my PhD, I also organized the CAMELYON16 challenge. From Jun to Nov 2016, I was a visiting researcher at Harvard University.
Qualcomm AI Research, Amsterdam, The Netherlands
Research updates:
- 23 May 2025: Mixture of Cache-Conditional Experts for Efficient Mobile Device Inference got accepted to TMLR 2025
- 09 May 2025: I gave an invited talk on "Efficient Deployment of LLMs on Edge Devices" at GHOST Day, Poznan, Poland
- 10 Mar 2025: This week, I delivered invited talks on LLM Efficiency at Apple and Cisco
- 09 Dec 2024: We will be demoing our Cache-MoE running efficiently on a smartphone at NeurIPS'24
- 05 Dec 2024: We bring Mixture of Experts (MoE) to mobile devices with limited available DRAM
- 26 Sep 2024: Check out our NeurIPS'24 and BMVC'24 papers on Efficient MoE and Multi-task learning
- 21 Sep 2023: Check out our NeurIPS'23 paper: Scalarization for Multi-Task and Multi-Domain Learning at Scale