Shunit Haviv Hakimi

Hi! I'm Shunit. I'm building new things at the intersection of AI, audio, and video. I previously worked as an Applied Scientist at Amazon, where I developed automatic recaps for Prime Video.

The albums in the background are some of my favorites. Create your own /wall-of-sound.

brag

/brag is a skill I built that turns your project into a short launch video. Check it out here GitHub.

BebopNet

BebopNet creates personalized jazz improvisations. Hear the samples, watch the video, or read the blog post GitHub.

Papers

BebopNet: Deep Neural Models for Personalized Jazz Improvisations

Shunit Haviv Hakimi, Nadav Bhonker, Ran El-Yaniv

ISMIR 2020 · Best Paper · demo · code
Group-Aware Reinforcement Learning for Output Diversity in Large Language Models

Oron Anschel, Alon Shoshan, Adam Botach, Shunit Haviv Hakimi, Asaf Gendler, Emanuel Ben Baruch, Nadav Bhonker, Igor Kviatkovsky, Manoj Aggarwal, Gerard Medioni

EMNLP 2025
Scene-VLM: Multimodal Video Scene Segmentation via Vision-Language Models

Nimrod Berman, Adam Botach, Emanuel Ben-Baruch, Shunit Haviv Hakimi, Asaf Gendler, Ilan Naiman, Erez Yosef, Igor Kviatkovsky

CVPR 2026