I'm exploring how to build more capable, efficient, and accessible AI for Indian languages. My current focus is Indic LLMs, including continued pretraining, tokenizer design, supervised fine-tuning, distillation, quantization, and model evaluation.
At Indian Institute of Science (IISc), I worked on diffusion models and GPU optimization as an AI/ML Research Intern.
I've also worked on deploying LLM and RAG systems for real-world applications. My broader interests include multimodal models and efficient inference.
I'm currently an AWS Student builder Group Lead at Amazon Web Services (AWS), and a Core committee member at Model United Nations.

