I’m working as an Applied Scientist-2 at Amazon with International Machine Learning team. I’ve helped in building and launching conversational shopping assistant for Amazon (India), both in the pre and post LLM era. I’ve built multiple NLU components (Intent/Planner, NER, Query Reformulation), Safeguard service, Next Question Suggestion service, Clarification and Product QnA tools for the assistant. I’ve also worked on personalization and forecasting problems in my first ~8 months at Amazon.

Previously, I have worked as a Machine Learning Scientist at Jio Haptik on fundamental Conversational-AI problems. I built the Intent Detection System for Haptik’s NLU Engine, which was 25% more accurate than their previous system, owning it from Research to Production.

I have authored research papers which have been accepted at top tier venues like ACL (Findings), EMNLP NLP-OSS workshop, EMNLP Insights workshop, EACL LT-EDI workshop and FIRE. I’ve also published 5 papers (3 orals (<10% acceptance rate) and 2 posters (<30% acceptance rate)) at Amazon’s internal ML conference (AMLC). See my resume for brief details on all external and internal papers.

I am also the creator of open source iNLTK library which provides out of the box support for various NLP tasks, for low resource 13 Indic Languages. The library has 175,000+ downloads, 800+ stars and 100+ forks on GitHub.

Prior to Jio Haptik, I worked at Goldman Sachs as a Software Engineer with the User Experience and Productivity team on Analytics for Desktop Assistant, which is firm-wide used productivity tool.

I have Masters in Computer Science with specialization in ML from Georgia Tech and Bachelor’s in Computer Science from PEC University of Technology.

I am interested in the applications of Machine Learning to solve problems which will impact millions and keep making my little open source contributions towards it.

Publications

Google Scholar

Education

Industry Experience

Open Source Contributions

	Natural Language Toolkit for Indic Languages (iNLTK) Star Fork Watch • iNLTK aims to provide out of the box support for various NLP tasks that an application developer might need for Indic languages • iNLTK provides Data Augmentation, Sentence Similarity, Sentence Encoding, Word Embedding, Tokenization and Text Generation utilities for low resource 13 Indic Languages • The library is backed by ULMFiT Language Models which I had trained using Fastai and Pytorch libraries, producing SOTA LM perplexity and Classification accuracy in 13 Indic Languages Appreciation for iNLTK • By Jeremy Howard, Sebastian Ruder on Twitter • Shared a lot by community on LinkedIn • iNLTK has 100,000+ Downloads on PyPi • Data Augmentation post about iNLTK was trending on LinkedIn • iNLTK was trending on GitHub in May, 2019 • Shared on Reddit, Facebook, Quora etc by the community
	Code with AI Star Fork Watch • Tool which predicts which techniques one should use to solve a competitive programming problem to get correct answer • Demo video on YouTube Appreciation for Code with AI • By Jeremy Howard on Twitter • By community on Codeforces • The tool has been used by 3000+ users
	NLP for Hindi Star Fork Watch • Contains SOTA Language models and Classifier for Hindi • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Sanskrit Star Fork Watch • Contains SOTA Language models and Classifier for Sanskrit • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Nepali Star Fork Watch • Contains SOTA Language models and Classifier for Nepali • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Tamil Star Fork Watch • Contains SOTA Language models and Classifier for Tamil • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Bengali Star Fork Watch • Contains SOTA Language models and Classifier for Bengali • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Punjabi Star Fork Watch • Contains SOTA Language models and Classifier for Punjabi • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Malayalam Star Fork Watch • Contains SOTA Language models and Classifier for Malayalam • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Odia Star Fork Watch • Contains SOTA Language models and Classifier for Odia • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]
	NLP for Gujarati Star Fork Watch • Contains SOTA Language models and Classifier for Gujarati • Pretrained Models available for download: TransformerXL, ULMFiT [ Code ] [ Results ] [ Dataset ] [ Embeddings projection ]

Honors & Awards

Skills and Courses

Last updated on 2021-10-03