Nitzan Hodos

prof_pic.jpg

I am a Ph.D. candidate at the Taub Faculty of Computer Science at Technion, under the supervision of Prof. Assaf Schuster.

Currently interning at Amazon Prime Video, focusing on using generative models for improved 3D reconstruction.

My research focuses on efficiency and computer vision — from efficient training and inference of LLMs and VLMs, to 3D reconstruction and video diffusion.

Publications

2026

  1. anchorkv.png
    AnchorKV: Anchor-Residual KV Cache Compression
    Malik Khalaf, Yara Shamshoum, Nitzan Hodos, Yuval Sieradzki, Assaf Schuster
    Under Review, 2026 arXiv
  2. macro.jpg
    MACRO: Training-free Multi-plane Attention for Closeup Render Optimization
    Nitzan Hodos, Roy Amoyal, Lior Fritz, Ianir Ideses, Sagie Benaim, Netalee Efrat
    NeurIPS 2026 (Spotlight) arXiv Code Project Page
  3. pamm.png
    QKV Projections Require a Fraction of Their Memory
    Malik Khalaf, Yara Shamshoum, Nitzan Hodos, Yuval Sieradzki, Assaf Schuster
    ICLR 2026 (Poster) arXiv Code

2025

  1. compact.png
    CompAct: Compressed Activations for Memory-Efficient LLM Training
    Yara Shamshoum*, Nitzan Hodos*, Yuval Sieradzki, Assaf Schuster
    NAACL 2025 (Oral) arXiv Code

2024

  1. dnc.png
    DNCs Require More Planning Steps
    Yara Shamshoum, Nitzan Hodos, Yuval Sieradzki, Assaf Schuster
    ICML 2024 (Poster) arXiv

2022

  1. cfnn.png
    Coin Flipping Neural Networks
    Yuval Sieradzki, Nitzan Hodos, Gal Yehuda, Assaf Schuster
    ICML 2022 (Poster) arXiv