Ayush Barik

About Me

I (wish to) revolutionize the world with cheap and powerful Generative AI. I primarily work at the intersection of Diffusion Modeling and Systems. I do systems research under Prof. Fan Lai and Prof. Francis Yan. I also do more theoretical research under Prof. Romit Choudhury. I have also collaborated with industry researchers from companies like Google and Amazon. In my personal time I like to watch anime/manga, read webnovels, play chess, and invest. I plan to pursue a PhD.

University of Illinois Urbana-Champaign
B.S. Computer Engineering (2023-2027)

News

(Feb 21, 2026) "RewardFlow: Generate Images by Optimizing What You Reward" was accepted to CVPR 2026!
(Sep 18, 2025) "Toward Engineering AGI: Benchmarking the Engineering Design Capabilities of LLMs" was accepted to NeurIPS 2025!
(Sep 12, 2024) Received the Illinois Engineering Achievement Scholarship!

Publications

[4] SoundWeaver: Semantic Warm-Starting for Text-to-Audio Diffusion Serving (INTERSPEECH 2026 under review)
Ayush Barik¹, Sofia Stoica¹, Nikhil Sarda², Arnav Kethana¹, Abhinav Khanduja¹, Muchen Xu¹, Fan Lai¹
¹UIUC, ²Assured Intelligence

[3] SparFusion: Efficient Diffusion Model Serving with Semantic Patch Parallelism (SIGMETRICS 2026 under review)
Yaqi Qiao¹, Ping He², Songrun Xie³, Ayush Barik¹, Chensong Zhang⁴, Zhengzhong Tu⁵, Fan Lai¹
¹UIUC, ²Vanderbilt University, ³HKUST, ⁴Nvidia, ⁵Texas A&M University

[2] RewardFlow: Generate Images by Optimizing What You Reward (CVPR 2026)
Onkar Kishor Susladkar¹, Dong-Hwan Jang¹, Tushar Prakash², Adheesh Sunil Juvekar¹, Vedant Shah¹, Ayush Barik¹, Nabeel Bashir¹, Muntasir Wahed¹, Ritish Shrirao¹, Ismini Lourentzou¹
¹UIUC, ²Sony

[1] Toward Engineering AGI: Benchmarking the Engineering Design Capabilities of LLMs (NeurIPS 2025)
Xingang Guo^1*†, Yaxin Li^1*, Xiangyi Kong^1*, Yilan Jiang^1*, Xiayu Zhao^1*, Zhihua Gong^1*, Yufan Zhang^1*, Daixuan Li¹, Tianle Sang¹, Beixiao Zhu¹, Gregory Jun¹, Yingbing Huang¹, Yiqi Liu¹, Yuqi Xue¹, Rahul Dev Kundu¹, Qi Jian Lim¹, Yizhou Zhao², Luke Alexander Granger¹, Mohamed Badr Younis¹, Darioush Keivan¹, Nippun Sabharwal¹, Shreyanka Sinha¹, Prakhar Agarwal¹, Kojo Vandyck¹, Hanlin Mai¹, Zichen Wang¹, Aditya Venkatesh¹, Ayush Barik¹, Jiankun Yang¹, Chongying Yue¹, Jingjie He¹, Libin Wang¹, Licheng Xu¹, Hao Chen¹, Jinwen Wang¹, Liujun Xu¹, Rushabh Shetty¹, Ziheng Guo¹, Dahui Song¹, Manvi Jha¹, Weijie Liang¹, Weiman Yan¹, Bryan Zhang¹, Sahil Bhandary Karnoor¹, Jialiang Zhang¹, Rutva Pandya¹, Xinyi Gong¹, Mithesh Ballae Ganesh¹, Feize Shi¹, Ruiling Xu¹, Yifan Zhang¹, Yanfeng Ouyang¹, Lianhui Qin³, Elyse Rosenbaum¹, Corey Snyder³, Peter Seiler⁴, Geir Dullerud¹, Xiaojia Shelly Zhang¹, Zuofu Cheng¹, Pavan Kumar Hanumolu¹, Jian Huang¹, Mayank Kulkarni⁵, Mahdi Namazifar⁵, Huan Zhang¹, Bin Hu^1†
¹UIUC, ²University of Pennsylvania, ³UC San Diego, ⁴University of Michigan, ⁵Amazon AGI

About Me

News

Ongoing Project(s)

Publications

Teaching & Service