Curriculum Vitae
Education
International Institute of Information Technology, Hyderabad
B.Tech in Computer Science and MS (by Research) in Computational Linguistics
- CGPA: 8.94/10.00 (Total); 9.28/10.00 (MS)
- Dean's Academic List: 2021–22 (Top 5%), 2022–23 (Top 10%), 2023–24 (Top 15%)
- Dean's Research List: 2024–2025 (Top 5%)
Publications
PrivacyBench: A Conversational Benchmark for Evaluating Privacy in Personalized AI
Preprint
InterChart: Benchmarking Visual Reasoning Across Decomposed Charts
AACL 2025 | Website
MapIQ: Evaluating Multimodal Large Language Models for Map QA
COLM 2025 | Paper
PRAISE: Enhancing Product Descriptions with LLM-Driven Structured Insights
ACL 2025 (System Demonstrations) | Website
MAPWise: Evaluating Vision-Language Models for Advanced Map Queries
NAACL 2025 (Oral), Nominated for Outstanding Paper | Website
Unraveling the Truth: Do VLMs really Understand Charts?
EMNLP 2024 | Website
Research Experience
Precog Research Group, IIIT Hyderabad
Advisor: Prof. Ponnurangam Kumaraguru
Collaborators: Prof. Saurav Prakash (IIT-M), Prof. Sheetal Kalyani (IIT-M), Prof. Jisun An (Indiana University)
- Investigating fairness constraints in Federated Learning (FL) setups to mitigate algorithmic bias across heterogeneous clients
- Optimizing the training of Binarized Neural Networks (BNNs) to enable efficient learning and adaptation on resource-constrained devices
- Evaluating privacy risks in personalized agents to detect secret leakage and ensure alignment with Contextual Integrity frameworks
Language Technologies Research Center (LTRC), IIIT Hyderabad
Advisor: Prof. Manish Shrivastava (MS Thesis Advisor)
- Explored lateral model compression strategies to test the limits of parameter reduction in large-scale architectures
- Conducted comparative analysis of diverse Large Language Models (LLMs) to map representational similarities and architectural divergences
- Current Research: Investigating non-autoregressive frameworks, like diffusion models, to align language generation more closely with human cognitive patterns
- Systematically diagnosing the root causes of performance degradation in long-context models to understand the limits of effective context utilization
Whitebox AI Safety Fellowship
Mentor: Kyle Reynoso
- Conducted mechanistic interpretability research to diagnose the phenomenon of emergent misalignment in Large Language Models (LLMs) using cross-encoders
- Analysed diverse tasks that might lead to misaligned behaviors in models
- Recognized as a Top 3 Project in the cohort
Cognitive Computation Group (UPenn) / CORAL Lab (ASU)
Advisors: Prof. Dan Roth (UPenn), Prof. Vivek Gupta (ASU)
Collaborators: Dr. Aparna Garimella (Adobe), Pritika Ramu (Adobe), Prof. Ross Maciejewski (ASU)
- Investigated the robustness and consistency of Vision-Language Models (VLMs) in interpreting complex visual information, specifically charts, graphs, and geospatial maps
- Benchmarked the robustness of VLM reasoning capabilities on different kinds of charts, identifying failure points
- Developed "MAPWise" and "MapIQ" benchmarks to evaluate multimodal models on advanced geospatial queries, counterfactual reasoning, and map-reading literacy
Software Engineering Research Center (SERC), IIIT Hyderabad
Advisor: Prof. Karthik Vaidhyanathan
- Investigated automated code summarization techniques to extract functional natural language descriptions from code snippets, benchmarking on the APPS dataset
- Explored the integration of Abstract Syntax Trees (ASTs) to enrich model inputs with hierarchical structural context, aiming to enhance the semantic accuracy of generated summaries
Industry Experience
Enterpret
Machine Learning Research Intern
- Enhancing the taxonomy retrieval system to extract actionable insights from large-scale customer feedback data
- Leveraging Large Language Models and Retrieval Augmented Generation to improve the granularity and semantic accuracy of feedback classification
Software Engineering Intern | Team: Google Cloud Storage (GCS) Orion
- Created a topology visualization tool for GCS Orion, enabling engineers to isolate and debug critical network issues across 100,000+ nodes
- Received Pre-Placement Offer (PPO) for Full-Time Software Engineer role
Software Engineering Intern | Team: Android PDF
- Designed and implemented the core image selection pipeline for the Android PDF viewer
- Received Return Internship Offer for Summer 2025 based on exceptional performance
STEP Intern | Team: Google Cloud Storage (GCS) Metering
- Built integrity checks for a high-throughput pipeline processing petabytes of Cloud usage data, ensuring billing accuracy for enterprise customers
- Converted to standard Software Engineering Intern (Return Offer) for Summer 2024
Honors and Awards
- India AI Fellowship (2025): Fellowship and aid to conduct research on LLM Similarity
- Amazon ML Challenge (2024): Secured 13th Rank nationally among ~75,000 participants (Top 0.02%)
- Google Generation Scholarship (2024): Awarded for academic excellence and leadership in technology
- JPMorgan Chase Quant Research Mentorship (2024): Selected for the specialized quantitative research track
- D. E. Shaw Ascend Educare (2022): Selected as a Fellow for the prestigious mentorship cohort as one of the youngest participants
Teaching and Leadership
Teaching Assistantships
Instructor: Prof. Manish Shrivastava
Instructor: Prof. Aditi Mukherjee
Instructor: Prof. Suryajith Chillara
Volunteership
Team Leader & Mentor, International Linguistics Olympiad (IOL)
Bansko, Bulgaria
- Selected to lead and mentor the Indian National Team at the global finals
- Trained students on advanced linguistic puzzle-solving techniques during the training camp
- Helped manage team strategy and logistics during the competition in Bulgaria
Lead Student Coordinator, Panini Linguistics Olympiad (PLO), India
- Orchestrated the national-level selection process for the Indian team, managing logistics for thousands of participants and coordinating with regional centers
- Helped conduct the Asia Pacific Linguistics Olympiad in India to select the Indian team
Club Coordinator, Open Source Developers' Group (OSDG), IIIT Hyderabad
- Organized an intra-college beginner hackathon attracting over 150 participants from the first and second years
- Conducted technical workshops on Linux systems, Open Source contributions, and Google Summer of Code (GSoC) mentorship