Jaisidh Singh
I am a master’s student at the University of Tübingen studying machine learning and a fellow at Zuse School ELIZA. I’m also a guest researcher at the Max Planck Institute for Intelligent Systems Tübingen advised by Antonio Orvieto.
Coming from an engineering background, my approach to machine learning research is one that seeks to better understand the “bells-and-whistles” we add to modern neural networks to make them more practically useful and robust.
My research interests are primarily aligned with but not limited to architecture design, reasoning, and efficient training of foundation models.
If you’re interested in collaborating or chatting about these topics, reach out to me via X or email. I like getting messages!
News
| Nov 05, 2025 | Presented our work on hypernetworks for efficient multi-modal stitching at EMNLP 2025. |
|---|---|
| Aug 20, 2025 | Work done during ELLIS research assistantship acceped at EMNLP 2025 (Main) |
| Jan 01, 2025 | Started a research assistantship at ELLIS Institute Tübingen |
Blog posts
| Nov 20, 2025 | Celebration is the secret |
|---|---|
| Sep 07, 2025 | Ultra-Scale Playbook vol-3 - DeepSpeed ZeRO |
| Sep 03, 2025 | Ultra-Scale Playbook vol-2 - Data Parallelism |
Selected publications
- AFMS(Almost) Free Modality Stitching of Foundation ModelsIn Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, 2025
- FMWNLearning the Power of “No”: Foundation Models with NegationsIn 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025