Amruta Parulekar

Amruta Parulekar

CS PhD at UIUC

Amruta Parulekar

Hi! I am Amruta Parulekar, a first year Computer Science PhD student at the University of Illinois, Urbana-Champaign. I am an Amazon AI PhD Fellow and am advised by Prof. Hari Sundaram. I will also be closely collaborating with Prof. Dilek Hakkani-Tur.

I completed my B.Tech. in Electrical Engineering, M.Tech. in Artificial Intelligence & Data Science, and Minor in Computer Science from the Indian Institute of Technology Bombay. There, I worked with Prof. Amit Sethi on using multimodal (primarily image) data for medical diagnostics. In my third year, I interned with Prof. Lueder Kahrs and used computer vision techniques for laryngeal surgery assistance. For my Master’s Thesis, I was advised by Prof. Preethi Jyothi, and worked on multilingual speech recognition and generation for low resource languages. I have been awarded two Undergraduate Research awards by IIT Bombay for these research contributions.

Research Interests

The problems that I am currently exploring are:

  • Improving the reasoning of LLMs on multimodal, multi-lingual or domain-specific tasks
  • Understanding LLM behaviour such as sycophancy and toxicity and improving user safety
  • User-centric conversational AI, especially multi-lingual and multi-modal usecases

I love working on real-world Artificial Intelligence applications. I also enjoy mixing modalities and have worked with speech, text, image as well as numerical data. I want to use my natural language processing, computer vision and reinforcement learning knowledge to make a difference where I can!

Additional Interests

I also enjoy teaching and mentoring, having been a teaching assistant for several courses at IIT Bombay including Foundations of Machine Learning and Automatic Speech Recognition. I have also been a department academic mentor for two years and received the best mentor award. I have served as a mentor for several AI/ML programs in IIT Bombay clubs.

I love travelling, hiking, and everything to do with nature. Cats, dogs and coffee happen to be great weaknesses :) In my free time, I enjoy art, photography and music. I play the keyboard and have tried my hand at every medium, from pencil sketching, water colors, to acryllics on canvas and even roads.

News!

Nov 2025My work at IIT Bombay as first author, LASER: An LLM-based ASR scoring and evaluation rubric was published in EMNLP 2025 (Main conference)!
Aug 2025Received the Amazon AI PhD Fellowship!
Jun 2025Admitted to the University of Illinois Urbana-Champaign for a Computer Science PhD!
May 2025My work at IIT Bombay as first author, AMPS: ASR with Multimodal Paraphrase Supervision was published in NAACL 2025 (Main conference)!
Apr 2025My work at IIT Bombay titled PathoGen-X: A Cross-Modal Genomic Feature Trans-Align Network for Enhanced Survival Prediction from Histopathology Images was published in IEEE ISBI 2025!
Mar 2025My work at IIT Bombay titled Transforming Breast Cancer Diagnosis: Towards Real- Time Ultrasound to Mammogram Conversion for Cost-Effective Diagnosis was accepted to the Ultrasonics Journal!
Aug 2024My work at IIT Bombay as first author, Parameter-efficient Adaptation of Multilingual Multimodal Models for Low-resource ASR was accepted to the Multilingual Representation Learning workshop of EMNLP 2024!
Jul 2024Scored 332/340 in the Graduate Records Exam (GRE) and 115/120 in the TOEFL!
May 2024At the 145th American Laryngological Association meet (COSM), Chicago, to present our poster on a computer vision pipeline for laryngoscopic image standardization!
Apr 2024Awarded the Best Mentor Award under the department academic mentorship program at IIT Bombay!
Feb 2024At the Bioimaging conference (BIOSTEC), Rome, to present our paper on a novel loss function to combine datasets with different label sets for improved nucleus segmentation and classification - Best Student Paper Award nomination!
Jan 2024My work at the University of Toronto as first author, A Computer Vision Pipeline for Laryngoscopic Image Standardization through Histogram Matching was accepted to the 145th ALA meet as a poster presentation!

Flag Counter