Shriya Atmakuri

Shriya Atmakuri

Skilled Applied AI Researcher & Data Scientist with background in Natural Language Processing and experience translating projects from user requirements to production-ready

JPMorgan Chase & Co.

University of Massachusetts Amherst

Biography

I am an AI & ML Associate, Sr. at the Machine Learning Center of Excellence in JPMorgan Chase & Co. working on news analytics and other exciting applied NLP research projects.

I have an MS in Computer Science student from the University of Massachusetts (UMass) Amherst. While doing my master’s, I’ve had the opportunity to work in the Information Extraction and Synthesis Lab headed by Prof.Andrew McCallum. I’ve also collaborated with Taesung Lee and Hessel Tuinhof at IBM Research. I interned for one summer at the Machine Learning Center of Excellence headed by Dr Lidia Mangu at JP Morgan Chase & Co. where I now work full-time.

I am passionate about delivering NLP and LLM-based solutions that satisfy users’ requirements, ensuring they effectively address their needs and enhance their overall experience.

My research interests are primarily linked to Natural Language Understanding. I am interested in what NLU sytems can tell us about human behavior and in what human cognition can teach us about building better NLU models. I am also interested in ethical, and explainable NLP and applying it towards social good.

Previously, I was a software engineer at Microsoft working on Azure and obtained my undergraduate degree in Computer Science from Manipal Institute of Technology, India.

In my free time, I enjoy reading and board gaming.

Interests
  • Natural Language Processing
  • Generative AI
  • Computational Linguistics
  • NLP for Social Science
  • Information Retrieval
  • AI for Good
Education
  • MS in Computer Science, 2022

    University of Massachusetts Amherst

  • BE in Computer Science, 2018

    Manipal Institute of Technology

Publications

(2022). Robustness of Explanation Methods for NLP Models. Workshop on Trustworthy Artificial Intelligence as a part of the ECML/PKDD 22.

PDF Cite Code

(2022). Word2Box: Capturing Set-Theoretic Semantics of Words using Box Embeddings. ACL 2022.

PDF Cite Code

Experience

 
 
 
 
 
JP Morgan Chase & Co
AI & ML Associate Sr.
JP Morgan Chase & Co
Jan 2023 – Present New York
 
 
 
 
 
JP Morgan Chase & Co
AI & Data Science Summer Associate
JP Morgan Chase & Co
Jun 2022 – Aug 2022 New York
 
 
 
 
 
Information Extraction and Synthesis Lab
Graduate Student Researcher
Feb 2021 – Aug 2021 Amherst
 
 
 
 
 
Microsoft
Software Engineer
Microsoft
Jul 2018 – Apr 2020 India

Projects

File Encryption System

File Encryption System

Minimal web page allowing users to upload files and encrypt them with AES.

Question Generation for Conversational Question Answering

Question Generation for Conversational Question Answering

Trained a GPT-2 based Question Generation Model on the QuAC dataset and experimented with using generated questions to enhance training of BERT-based Question Answering Model

Supervised Clustering for Fact-Checked Claim Retrieval

Supervised Clustering for Fact-Checked Claim Retrieval

Addressed the task of verified claim retrieval - linking an unseen claim to previously-seen claims which have already been fact-checked. Used supervised clustering approach which defined a similarity measure between claims and facts using BERT-based representations.

Emotions on Reddit

Emotions on Reddit

Emotion classification on Reddit comments using BERT-base finetuned on GoEmotions dataset. Studied relationships between emotions expressed and upvotes given.