Work Expereince
Erik Jonsson School of Engineering and Computer Science, UT Dallas September 2021 - December 2022
- Grader for Graduate level Data Structures course (CS 5343) Fall 2022
- Teaching assistant for Computer Programming I (CS 1337) Spring 2022
- Teaching Assistant and Grader for Graduate level Discrete Structures course (CS 5333) Fall 2021
AI-NLP-ML Lab, IIT Patna January 2021 - June 2021
- Designed mental health disorder identification systems, with a goal of creating a virtual Cognitive Behavioral Therapy bot
- Curated a large-scale dataset consisting of dyadic conversations for multiple mental illnesses
- To identify and distinguish different mental disorders from the counselling conversations, we developed a dual attention (self and cross) based DNN classifier built on top of the BERT
Dr. Swati Agarwal's Lab, BITS Pilani, Goa Campus Jan 2021 - June 2021
- Developed a Hierarchical Attention Neural network for identifying ironic tweets, by leveraging semantic features likes POS tagging and sentiment flow shifts
- Developed an attention-based BiLSTM-CNN framework for the purpose of generating diverse extended summaries of scholarly articles
AI-NLP-ML Lab, IIT Patna Apr 2020 - Aug 2020
Participated & developed systems for 3 shared task of Scholarly Document Processing workshop at EMNLP 2020
- CL-SciSumm: The 6th Computational Linguistics Scientific Document Summarization Shared Task
- CL-LaySumm: The 1st Computational Linguistics Lay Summary Challenge Shared Task
- LongSumm: Shared Task on Generating Long Summaries for Scientific Documents
Dr. Ambika Pawar's Lab, SIT Pune Feb 2020 - July 2020
- Defined a new task of, tweets reporting abuse classification, and constructed a small scale dataset related to the online abuse reporting. A detailed comparison with existing supervised models are explored.
AI-NLP-ML Lab, IIT Patna Dec 2019 - Jan 2020
- Developed a novel Multi-view clustering framework (MV-M-SDS) for solving the task of Scientific Document Summarization by considering both syntactic and semantic views
- We set a new state-of-the-art performance on two SDS shared task datasets(CL SciSumm 2016 & 2017) related to computational linguistic domain
AI-NLP-ML Lab, IIT Patna May 2019 - Aug 2019
- Worked on developing Multi Modal Named Entity Recognization (NER), which uses both Visual features and texual (word level & character level) embeddings for better NER accuracy.
- Achieved state-of-the-art performances on the large-scale multimodal tweet dataset
- Performed a detailed ablation study with combinations of text, image and url based modalities with attention mechanisms.