Experience

 
 
 
 
 
Feb 2025 – Present
Netflix Research | Los Gatos, CA

Research Scientist

Netlfix Research

Large Multimodal Language Model Research. Reward Modelling, Contrastive Learning, Finetuning for Agents
 
 
 
 
 
May 2024 – Dec 2024
Netflix, Localization Team | Los Gatos, CA

Research Scientist Intern

Netlfix Research

Working on multimodal contrastive learning for fine-grained audio-visual sync estimation.
 
 
 
 
 
May 2023 – Dec 2023
Yahoo, Visual Intelligence Team | San Francisco

Research Scientist Intern

Yahoo Research

Working as Research intern in visual intelligence team to solve multi-modal retrieval problem.
 
 
 
 
 
May 2022 – Dec 2022
Adobe, Media Intelligence Lab | San Jose, CA

Research Scientist Intern

Adobe Research

Designed novel pre-training strategies and transformer architectures for large scale vision-language contrastive pretraining on 100M scale datasets. Evaluation on downstream tasks such as zero-shot image classification, text-image retrieval, object-detection etc.
 
 
 
 
 
Jun 2020 – Present
Buffalo

Research Assistant

Department of Computer Science, University at Buffalo

I am Graduate Research Assitant (SUNY) at CUBS lab at University at Buffalo, State University of New York. I am working with Prof. Srirangaraj (Ranga) Setlur and Prof. Venu Govindaraju on an NSF funded project called Made@UB ML toolkit. ML Toolkit is an easy-to-use GUI based application, developed to reduce the time it takes to build prototypes for ML models as well as experiment with various feature extraction methods available.
 
 
 
 
 
Jul 2019 – Dec 2020
Pune

Software Engineer - Machine Learning

Persistent Systems

At Persistent Systems we build software that drives the business of our customers; enterprises and software product companies with software at the core of their digital transformation.

Recent Publications

Ph.D. dissertation, State University of New York at Buffalo, 2025

Recent Blogs

Google recently released the Gemma 3n models — E4B and E2B. The models are packed with novel components and features from PLEs, ASR and …

Typically, preference alignment in large language models (LLMs) requires a reference model and a warm-up phase of supervised …

LLMs are typically trained on fixed-length sequences, leading to performance degradation when dealing with longer texts due to …

The article explores Grouped Query Attention (GQA), an efficient pre-training strategy for large language models (LLMs) like LLaMA-2 …

Fine-tuning large pre-trained models is computationally challenging, often involving adjustment of millions of parameters. This …

Recent & Upcoming Talks

Lab’s weekly paper presentation | CVPR paper presented at regular lab meeting.

Azure for App Developers

Facebook Developer Cirlce Indore | F8 Indore Meetup | Pytorch: Deep Learning Framework

Facebook Developer Cirlce Indore | Women in Tech.

Getting started with opensource and creating your first VR app