Devang Kulshreshtha
💻

kulshrde@amazon.com

About

Hi! I'm Devang, an Applied Scientist II at Amazon AWS in NYC, USA. My research interests include AI safety, guardrails, detection of prompt attacks, red teaming, and explainable AI.

At Amazon I led and delivered the AWS Bedrock Guardrails Prompt Attack feature. I also delivered science contributions to multiple AWS launches such as HealthScribe and Lex, working on summarization and Automatic Speech Recognition.

I completed my MSc at the MILA Lab and McGill University with Prof. Siva Reddy. There, I interned with Korbit AI on question-answer generation from open educational resources like Wikipedia.

Education

MSc, Computer Science
Sept. 2020 -- June 2022

McGill University, Montreal, Canada

BTech, Computer Science
July 2014 -- May 2018

Indian Institute of Technology, India

Work Experience

July 2022 - Present, Amazon USA, Applied Scientist II
Jul 2021 - Oct 2021, Amazon UK, Applied Science Intern
September 2020 - June 2020, Korbit.AI, Research Intern
September 2018 - August 2020, Amazon India, Software Engineer
May 2017 - Jul 2017, Amazon India, SDE Intern
May 2016 - Jul 2016 , CFILT Lab IIT Bombay, Research Intern

Publications

Sequential editing for lifelong training of speech recognition models

Devang Kulshreshtha, Saket Dingliwal, Brady Houston, Nikolaos Pappas, Srikanth Ronanki Paper

INTERSPEECH 2024

Generalized zero-shot audio-to-intent classification

Veera Raghavendra Elluru, Devang Kulshreshtha, Rohit Paturi, Sravan Bodapati, Srikanth Ronanki Paper

IEEE ASRU 2023

Retrieve and copy: Scaling asr personalization to large catalogs

Sai Muralidhar Jayanthi, Devang Kulshreshtha, Saket Dingliwal, Srikanth Ronanki, Sravan Bodapati Paper

EMNLP 2023 (Industry Track)

Mask the bias: Improving domain-adaptive generalization of CTC-based ASR with internal language model estimation

Nilaksh Das, Monica Sunkara, Sravan Bodapati, Jinglun Cai, Devang Kulshreshtha, Jeff Farris, Katrin Kirchhoff Paper

ICASSP 2023

Multilingual contextual adapters to improve custom word recognition in low-resource languages

Devang Kulshreshtha, Saket Dingliwal, Brady Houston, Sravan Bodapati Paper

INTERSPEECH 2023

Few-shot question generation for personalized feedback in intelligent tutoring systems

Devang Kulshreshtha, Muhammad Shayan, Robert Belfer, Siva Reddy, Iulian Vlad Serban, Ekaterina Kochmar Paper

IJCAI-PAIS 2022

Back-Training excels Self-Training at Unsupervised Domain Adaptation of Question Generation and Passage Retrieval

Devang Kulshreshtha, Robert Belfer, Iulian Vlad Serban, Siva Reddy Paper Code

EMNLP 2021

How emotional are you? Neural architectures for emotion intensity prediction in microblogs

Devang Kulshreshtha, Pranav Goel, Anil Kumar Singh Paper Slides Code

COLING 2018

NLPRL-IITBHU at SemEval-2018 task 3: Combining linguistic features and emoji pre-trained CNN for irony detection in tweets

Harsh Rangwani, Devang Kulshreshtha, Anil Kumar Singh Paper

SemEval @ NAACL 2018

Feature Augmented Deep Neural Networks for Collaborative Filtering

Devang Kulshreshtha Paper Code

IJCAI 2017 Workshop on AI Applications in E-commerce

Prayas at emoint 2017: An ensemble of deep neural architectures for emotion intensity prediction in tweets

Pranav Goel, Devang Kulshreshtha, Prayas Jain, Kaushal Kumar Shukla Paper Slides

8th WASSA Workshop at EMNLP 2017

Patents

Automated evaluation of evidence mapping models

Devang Kulshreshtha, Saket Dingliwal, Sravan Babu Bodapati, Katrin Kirchhoff, Sarthak Handa URL

US Patent [18344739]

Medical Conversation Summarization Style Intelligence

Aparna Elangovan, Lei Xu, Devang Kulshreshtha, Sravan Babu Bodapati, Katrin Kirchhoff, Sarthak Handa URL

US Patent [18339749]

Academic Services

Reviewing: INTERSPEECH 2025, ARR 2024, INTERSPEECH 2024, ARR 2022

Talks:

Teaching: IIT Varanasi: Introduction to Artificial Intelligence (Spring 2018), Introduction to Programming in C (Fall 2016)

Devang Kulshreshtha 💻