Hi! I am a fourth-year PhD candidate at the Language Technologies Institute at Carnegie Mellon University. I am advised by Graham Neubig and have wonderful friends and collaborators at Neulab :)

Broadly, I am passionate about making AI genuinely useful for real-world applications and a diverse set of users. A lot of my research has drawn inspiration from what people actually need, but these systems lack. I am always excited to explore new directions which serves this grander goal, but my work thus far has focused on three main areas:

(1) LLM/Multimodal Evaluation: Evaluation is what truly drives progress, and it should be grounded in real-world needs and use-cases. I've built benchmarks and metrics that draw inspiration from what we actually use (or could use) these models for — culturally localizing images for domains like advertising and education (image transcreation, Best Paper, EMNLP 2024); measuring speech understanding across a broad, long-tail set of languages (FLEURS, Best Paper, SLT 2022); and handling code-mixed inputs, a phenomenon common in multilingual communities (GLUECoS).

(2) Models/Methodological Interventions: I probe and improve model representations for broader linguistic and cultural coverage — MuRIL, Pangea and Cultural Pangea, and inference-time steering for controllable, culturally aware generation.

(3) Data Selection & Synthetic Data: Data has consistently been a bottleneck in my research — whether working with the long-tail of languages and cultures, or with applications like advertising and education where much of the relevant data is copyrighted. I've worked on data-efficient fine-tuning — labeling strategies under a fixed budget (DeMuX) — and my in-progress/proposed work builds synthetic data pipelines to improve diversity sampling for text-to-image models and train multimodal generative models for applications like advertising and education.

I've been fortunate to have my work recognized through fellowships and awards including MIT EECS Rising Star, Rising Star in AI (UMich), BITS 30 Under 30 (Research), CMU Waibel Presidential Fellowship, and two Best Paper Awards at EMNLP 2024 and SLT 2022.

I'm deeply grateful to the brilliant researchers whose mentorship has shaped my growth: Graham Neubig (CMU), Partha Talukdar (Google DeepMind), Sebastian Ruder (Google DeepMind), Alexis Conneau (Google DeepMind), Sunayana Sitaram (Microsoft Research), Monojit Choudhury (Microsoft Research), and Dr. Sreejith V (BITS Pilani).

For more information, check out my CV or reach out via email :)