Research

My research lie in Natural Language Processing (NLP), with a focus on neuro-symbolic approaches, improving interpretability and alignment in large language models (LLMs), and meta-reasoning in LLMs.

Current Projects

NeuSyM-Meta: Neuro-Symbolic Meta-Reasoning for Self-Aware Logical Evaluation in Large Language Models

Exploring neuro-symbolic methods for meta-reasoning that will enable large language models (LLMs) to evaluate and improve their own reasoning processes.

AI Misalignment and Scheming Evaluation Framework

Studying AI misalignment with a focus on AI scheming, and developing benchmarks to evaluate and detect covert deceptive behaviors in advanced AI systems.

Recent Projects (2024-2025)

• Assessing Algorithmic Bias in Language-Based Depression Detection - Spring 2025 (Paper Accepted at BHI 2025 Conference)
  Investigating gender and racial disparities in DNN vs LLM approaches for depression detection, with fairness-aware mitigation strategies

• Structured Reasoning with LLMs for Question Answering over Tabular Data - Spring 2025
  SemEval-2025 Task 8 system using hybrid LLM strategies including retrieval-augmented generation and column-aware filtering

• Neuro-Symbolic Approach to Depression Detection - Fall 2024
  Integrating rule-based systems with Neural models such as Mental-RoBERTa for linguistic-based depression classification

• Detection of Insomnia from Clinical Notes - Spring 2025
  SMM4H-HeaRD 2025 shared task system using transformer models and structured metadata for automated insomnia detection from MIMIC-III clinical notes

• SmartMeet – AI Meeting Summarizer - Fall 2024
  Automated meeting summarization tool with key discussion point extraction, personalized to-do lists, and JIRA ticket integration