Komal Mathur

komlixmathur@gmail.com

I am a Junior studying Computer Science Engineering at Manipal Institute of Technology, India.

My academic journey has led me to develop a strong interest in Natural Language Processing and Generative AI. My current research interests are Fairness and Bias mitigation in ML Models, and Trustworthy and Explainable AI. Through internships at Info Edge and Samsung R&D Institute, I've gained practical experience in developing GenAI pipelines and working on LLM tokenization for smart home events. I'm also involved in multimodal AI research under Dr. Nisha Shetty, focusing on multimodal deception detection.

My goal is to continue exploring innovative use cases of AI, while also maintaining a check on correctness of content, and developing applications aimed at social good.

Experience

Generative AI Intern

Info Edge, Naukri.com
  • Building a robust MCQ Question Generation pipeline using GPT-4, ensuring complete automation through integration with MySQL.
  • Reduced token usage and optimized prompts, ensuring cost-effectiveness and scalability.
May 2024 - Present

Samsung PRISM Intern

Samsung R&D Institute India
  • Pioneering Effective LLM Tokenization for Smart Home Events.
  • Focused on techniques for integrating RAG to reduce the need for frequent LLM fine-tuning, enabling real-time processing of smart home events.
April 2024 - Present

Research Society Manipal

AI Senior Member
  • Working on Multimodal Deception detection under Dr. Nisha Shetty.
  • Proposed a novel architecture using the feature embeddings from multiple modalities and combining using latent space.
December 2022 - Present

Research Intern

Department of ICT
  • Working on Multimodal Deception detection under Dr. Nisha Shetty.
  • Proposed a novel architecture using the feature embeddings from multiple modalities and combining using latent space.
September 2023

Education

Manipal Institute of Technology

Bachelor of Computer Science

GPA: 9.95/10

Rank: 1 out of 2684
Diamond Jubilee Award
2x Achiever's Scholarship Recipient
2022 - 2026

Lotus Valley International School

Advanced Mathematics + Physics + Chemistry + CS

CBSE 12th: 98.6% (School Topper)

CBSE 10th: 96.6%

2008 - 2022

Publications

1. Bangla Sentiment Analysis using Weighted and Majority Voted Fine-Tuned Transformers

[Link]
Seth, P., Goel, R., Mathur, K., Vemulapalli, S. (2023)
Accepted at the 1st workshop on Bangla Langauge Processing at EMNLP 2023
This paper aims to promote work on Bangla Sentiment Analysis while identifying the polarity of social media content by determining whether the sentiment expressed in the text is Positive, Negative, or Neutral. Our approach consists of experimenting and finetuning various multilingual and pre-trained BERT-based models on our downstream tasks and using a Majority Voting and Weighted ensemble model that outperforms individual baseline model scores.

Projects

Hey Taxii App

[GitHub]

Developed a ride-share web application using MySQL, SQLAlchemy, Python, and HTML, equipped with the Google Maps API.

Hey Taxii App Photo

Student Performance Indicator

[GitHub]

Developed an AI-driven performance predictor using modular programming in Python, complete with deployment using Microsoft Azure Services (MLOps) and GitHub Actions.

Student Performance Indicator Photo

Custom-Emoji-Face-Generator

[GitHub]

A Java-based application designed to create customised emoji faces using Object-Oriented Programming (OOP) concepts and JavaFX, allowing users to dynamically generate emoji faces by selecting different face shapes, eye shapes, mouth shapes, and customizing colors.

Project 3 Photo

Skills

Programming Languages & Tools
Research Interests (currently exploring)
  • Natural Language Processing
  • Generative AI
  • Hallucinations in LLMs and Bias mitigation
  • Fairness and Privacy preservation in ML models
  • Trustworthy AI
  • Explainable AI
  • Causal Inference

Awards & Achievements

  • Ranked among Top 50 applicants to the 5th Cohort of Women in Engineering supported by Google, and offered 100% scholarship and a reward of Rs. 100,000
  • Received the merit-based Achiever’s Scholarship awarded to top 5 students in each batch for academic excellence in the 2nd year.
  • Awarded the Director’s Diamond Jubilee Award for being ranked 1/2684 students in the first year.
  • Received the merit-based Achiever’s Scholarship worth Rs.1,01,800 for academic excellence in the first year.

More

  • I love going on road trips and camping in nature. Here are some of my favourite shots
  • Last year, I clocked almost 100,000 minutes on Spotify (2.3 months of non-stop music)
  • My favourite song is 'Time' by Pink Floyd