username

Anshul Singh

👩🏻‍💻 Large Language Models | 🧠 Reasoning | 🌍 Multimodal & Multilingual AI
Bangalore, India
I’ve recently graduated from Panjab University, Chandigarh. I’m currently working as a Research Associate at the IACV Lab, IISc Bangalore, advised by Prof. Soma Biswas. Previously, I was part of the Language Technology Group @ University of Hamburg, a MITACS Globalink Research Intern, and served as ML/AI Lead at Google Developer Student Clubs.

My work focuses on Large Language Models (LLMs), multimodal systems, and document understanding and retrieval. I’m particularly interested in the reasoning capabilities of LLMs—especially in visual reasoning, multilingual and multimodal settings.

🔍 Currently: Researching multi-hop reasoning in multimodal models, particularly over tabular data.
📚 I love reading arXiv papers and tinkering with small experiments. Always open to collaboration and research discussions!

🌿 INTJ | I enjoy traveling, reading, and minimalist living. Happy to connect and become friends!

Writings

📘 Pruning and Quantizing the Neural Nets
📗 Language Models as Operating System
📕 Cuda and GPUs

Timeline



Research Interests and Inspo ✨

Reasoning in Large Language Models and Interpretability

I am interested in how LLMs reason and plan and how we can make these processes interpretable.

Multimodality and Retrieval

Multimodality - Vision & Text; how integrating vision and language enhances the retrieval process, especially for low-resource languages.

Visual Stream ✨

Gallery image: ReinGallery image: Me3Gallery image: Me4Gallery image: Dal2