Anand Subramanian

Anand Subramanian

MComp (CS) student at NUS

National University of Singapore

Biography

I’m a Master’s student specializing in Computer Science at the National University of Singapore. My current research interests lie in Large Language Models, particularly their application in clinical, scientific, and biomedical texts to enhance information extraction and question-answering.

Interests
  • Deep Learning
  • Natural Language Processing
Education
  • MComp in Computer Science, 2023

    National University of Singapore

  • BEng in Electronics and Communications Engineering, 2019

    Anna University (from SSN College of Engineering)

Experience

 
 
 
 
 
Research Engineer Intern
May 2023 – September 2023 Singapore
  • Developed Clinical Question Answering (QA) models using LLMs with Low-Rank Adapters. Created a novel benchmark evaluation suite comprising 22 diverse datasets to evaluate Large Language Models (LLMs) in zero-shot, few-shot and finetuned settings on QA tasks.

 

  • Published a paper at the ImageCLEF MediQA 2023 Workshop, which involved LLMs for generating medical notes from doctor-patient conversations using novel synthetic data augmentation modules.

 

  • Developed solutions for database schema generation using LLMs and In-Context Learning. Implemented evaluation modules for measuring the performance of the developed solution.

 

  • Full-time intern between May-August, Part-time intern between September and November
 
 
 
 
 
Research Engineer
September 2019 – July 2022 India
  • Researched and developed various deep-learning and NLP-based solutions for solving problems in the healthcare space. Worked on systems for performing automated clinical coding utilizing knowledge graphs and text-mining systems.

 

  • Developed novel models for OCR-robust section identification, negated entity identification and punctuation restoration for Clinical Electronic Health Records (EHRs).

 

  • Represented the company by publishing and presenting our work at several prestigious international conferences and workshops (EndoCV Workshop @ ISBI-2020, EDNIL @ FIRE-2020, ICON-2021, CODS-COMAD 2023).

Contact

If you think your research interests might align with my work and you’re interested in collaborating, or if you have any general questions, don’t hesitate to get in touch with me!