PhD Student
CMU · MLD

Abhinav Java

I develop ways for language models to efficiently find, use, and learn from large unstructured data sources. My work spans retrieval-augmented generation, efficient post-training, and vision-language models.

Before CMU, I was a Research Fellow at Microsoft Research India (with Amit Sharma and Nagarajan Natarajan). Even before that, I worked as a Research Associate at Adobe Research, and did an undergrad in CS at Delhi Technological University where I got to collaborate with the MIT Media Lab.

Email Scholar GitHub Twitter CV

I'm responsive over email and happy to chat about research, collaborations, and grad school applications.

Affiliations

2026 – now

PhD Student, Carnegie Mellon University, MLD

2024 – 26

Research Fellow, Microsoft Research India

2022 – 24

Research Associate, Adobe Research (MDSR)

2018 – 22

B.Tech Computer Engineering, Delhi Technological University

Selected Research

* equal contribution

CVPR 2026 · Oral

Understanding Task Transfer in VLMs

Bhuvan Sachdeva*, Karan Uppal*, Abhinav Java*, Vineeth N Balasubramanian

arXiv Project

ICLR 2026

FrugalRAG: Less is More in RL Finetuning for Multi-Hop Question Answering

Abhinav Java, Srivathsan Koundinyan, Nagarajan Natarajan, Amit Sharma

arXiv Code

ICLR 2026

Characterizing Deep Research: A Benchmark and Formal Definition

Abhinav Java*, Ashmit Khandelwal*, Sukruta Midigeshi*, Aaron Halfaker, Amit Deshpande, Navin Goyal, Ankur Gupta, Nagarajan Natarajan, Amit Sharma

arXiv Dataset

NAACL 2025

Towards Operationalizing Right to Data Protection

Abhinav Java*, Simra Shahid*, Chirag Agarwal

arXiv Code

WACV 2025

ReEdit: Multimodal Exemplar-Based Image Editing

Ashutosh Srivastava*, Tarun Menta*, Abhinav Java*, Avadhoot Jadhav, Silky Singh, Surgan Jandial, Balaji Krishnamurthy

arXiv Project

AAAI 2024

All Should Be Equal in the Eyes of LMs: Counterfactually Aware Fair Text Generation

Pragyan Banerjee*, Abhinav Java*, Surgan Jandial*, Simra Shahid*, Shaz Furniturewala, Balaji Krishnamurthy, Sumit Bhatia

arXiv Code

WACV 2023

One Shot Doc Snippet Detection: Powering Search in Document Beyond Text

Abhinav Java, Shripad Deshmukh, Milan Aggarwal, Surgan Jandial, Mausoom Sarkar, Balaji Krishnamurthy

Paper

ECCV 2022

Learning to Censor by Noisy Sampling

Ayush Chopra, Abhinav Java, Abhishek Singh, Vivek Sharma, Ramesh Raskar

arXiv Code

NeurIPS Workshop 2022

Rethinking Neural Networks With Benford's Law

Surya Kant Sahu, Abhinav Java, Arshad Shaikh, Yannic Kilcher

arXiv Code

Updates

Oct 25

VLM Task Transferability accepted to CVPR 2026 as an Oral!

Oct 25

FrugalRAG and LiveDRBench accepted to ICLR 2026.

Oct 25

FrugalRAG at NeurIPS 2025 Efficient Reasoning Workshop. Code.

Sep 25

Fireside Chat on Deep Research at Plutos Dev.

Sep 25

LiveDRBench on Hugging Face + NeurIPS 2025 SEA Workshop.

Aug 25

Career talk at DTU AI Summer School.

Aug 25

Task Transfer analysis at NeurIPS 2025 UniReps Workshop.

Jul 25

Presented FrugalRAG at ICML 2025 in Vancouver.

Jun 25

FrugalRAG at ICML EsFoMo Workshop. Preprint.

Feb 25

Unlearnable Text Datasets at NAACL 2025 (Main). Code.

Nov 24

ReEdit at WACV 2025 and ECCV-AI4VA 2024.

Sep 24

Thinking Fair and Slow at EMNLP 2024 (Main).

Aug 24

Joined Microsoft Research India as a Pre-Doctoral Research Fellow.

Beyond the lab

Football

I've played my whole life. Currently washed but used represent my high school and college teams.

Mentorship

Had the best time mentoring several research interns at Adobe.

Reviewing

CVPR '26, ICLR '26, CVPR '25, KDD '24, CVPR '24, U&ME@ECCV '24, WACV '23, ML4PS@NeurIPS '22, ECCV '22

How it started

My first NeurIPS workshop paper started as a question about neural network statistics, and a serendipitous collaboration with Yannic Kilcher.