Aug 2021 – Dec 2021 · 4 mos · Seattle, Washington, United States · Remote

• Worked on a zero-shot approach to improve passage retrieval.
• Internship paper "Improving Passage Retrieval with Zero-Shot Question Generation"(https://arxiv.org/abs/2204.07496) accepted at EMNLP 2022 conference.

Mar 2021 – Jul 2021 · 4 mos · Palo Alto, California, United States · Remote

• Research on domain-adaptation methods without finetuning for neural machine translation.
• Worked on retriever-based approaches for machine translation.
• Paper accepted at EACL Findings 2023.
• Filed an USPTO patent on the work jointly with Cuong Hoang, Prashant Mathur, and Marcello Federico.

Jun 2020 – Sep 2020 · 3 mos · Santa Clara, California, United States · Remote

• Worked on sequence-to-sequence models for large-scale language model pre-training and retrieval-augmented language models.
• Research on improving pre-training methods for generative language models such as T5; provided technical support to Amazon Web Services (AWS) team to implement the T5 model (see the blog below).
• Implemented learnable text-retrieval approaches and trained them on Wikipedia.
• Work involved training large-scale language models in a distributed setting consisting on 500+ GPUs.
• Publication accepted as oral in ACL 2021: End-to-End Training of Neural Retrievers for Open-Domain Question Answering (https://arxiv.org/abs/2101.00408)

Sep 2019 – Dec 2023 · 4 yrs 3 mos · Montreal, Quebec, Canada · On-site

• NLP Research

Feb 2018 – Aug 2019 · 1 yr 6 mos · Greater Pittsburgh Area

Aug 2017 – Dec 2017 · 4 mos · Greater Pittsburgh Area

• Convex Optimization (10-725)
• Prof. Aarti Singh, Prof. Pradeep Ravikumar, Machine Learning Department, School of Computer Science

May 2017 – Aug 2017 · 3 mos · Greater Pittsburgh Area

• Knowledge Graph Construction (Sailing Lab)

Oct 2014 – Jun 2016 · 1 yr 8 mos · Hyderabad Area, India

• Worked on customer reviews mining which included unsupervised keyword generation (phrasing), sentence segmentation, word embeddings, word sense disambiguation system, sentiment and aspect analysis.
• Worked on writing expert reviews scrapers, analysing and scoring mobile phone specifications and building a content based recommendation engine for mobile phones.
• Filed 2 patents on sentiment analysis, scoring and a novel way of feature extraction in text classification.

Nov 2013 – Sep 2014 · 10 mos · Hyderabad Area, India

• Worked on bayesian topic modelling, matrix factorisation algorithms to infer themes and draw actionable insights from reviews, surveys and user complaints.
• Worked on text classification approaches for document categorisation.

May 2012 – Jun 2012 · 1 mo

• Worked on Kernel Methods in Machine Learning.

May 2011 – Jul 2011 · 2 mos

• Worked on Image Processing