Large Language Models & Agentic AI
Tool-using agents, RAG, multi-agent systems, evaluation, structured reasoning, and reliable decision-making.
ASSISTANT PROFESSOR · COMPUTER SCIENCE · IIT BHUBANESWAR
Shreya Ghosh
Building intelligent systems that reason about language, space, uncertainty, and human preferences.
My research spans Natural Language Processing, Large Language Models, Agentic AI, Travel Intelligence, and Geospatial AI, with an emphasis on reliable reasoning in complex real-world settings.
ABOUT
AI grounded in the complexity of the real world.
I am an Assistant Professor in the School of Electrical and Computer Sciences at IIT Bhubaneswar. Before joining IIT Bhubaneswar, I was a postdoctoral scholar at The Pennsylvania State University. I received my PhD from IIT Kharagpur.
I work on NLP and AI systems that must operate beyond clean benchmark settings: travel agents that respect constraints and disruptions, multimodal systems that reason geographically, and language models that retrieve, plan, and make decisions using external evidence. A recurring theme across my work is the design of datasets, benchmarks, and end-to-end systems for realistic evaluation.
RESEARCH
Research themes
My work connects language intelligence with structured reasoning, real-world constraints, multimodal evidence, and spatial context.
Travel Planning & Intelligence
Constraint-aware itinerary generation, review-grounded personalization, uncertainty, robustness, and disruption-aware planning.
Geospatial AI & Mobility
Geolocalization, mobility analytics, multimodal spatial reasoning, urban intelligence, and location-aware systems.
NLP & Misinformation
Early misinformation detection, social-media reasoning, multilingual NLP, and human-centered language technologies.
SELECTED RESEARCH
Research themes and ongoing work
Each theme brings together related papers, collaborators, and representative results. Open a card for more detail.
Open a research themeSelected & recent work
A selection of recent papers across language models, travel planning, multimodal reasoning, education, and misinformation.
Selected papersScroll to browse
TEACHING
Courses & professional teaching
I teach core computer science and AI courses at IIT Bhubaneswar, alongside professional programs on contemporary generative and agentic AI.
PEOPLE
Students & collaborations
Research in the group is strongly student-driven, with collaborations across academia and industry.
PhDDoctoral researchers
- Priyanshu Karmakar
- Divyansh Srivastava
- Rousan
MTech / MSRMaster's researchers
- Soutrik Das
- Anushka Saxena
- Sayan Dey
- Arindita Pal
BTechUndergraduate researchers
- Soumyabrata Chaudhuri 2025
- Vijay Sai Vorru 2026
- Suprit Naik 2026
This list will be updated as new students join.
ACROSS ACADEMIA & INDUSTRY
Collaborative research
Current and recent work includes collaborations with researchers at Microsoft and academic partners in India and abroad.
INDUSTRY
Microsoft
Manish Gupta · Shubhojit Mallick
ACADEMIC
University of Melbourne
ACADEMIC
IIT Kharagpur
ACADEMIC
MBZUAI
Recent work & highlights
A quick view of recent publications and milestones. Scroll for more.
2025–2026Latest first
Three full papers accepted to EMNLP 2026 Findings
TripPulse, UTP-Bench, and GeoAgent.
TripTide accepted to ACL 2026 Findings
Adaptive travel planning under disruptions.
IntelliCode at EACL 2026 System Demonstrations
A multi-agent LLM tutoring system with centralized learner modeling.
mmJEE-Eval in IJCNLP-AACL 2025 Findings
Bilingual multimodal scientific reasoning in English and Hindi.
Cultural competence in vision-language models at ICCVW 2025
Multimodal story generation as a testbed for culturally grounded model behavior.
TripCraft at ACL 2025 Main Track
Spatio-temporally fine-grained travel planning with real-world data and continuous evaluation.
CONTACT
Interested in working together?
I welcome research conversations with motivated students and collaborators working on NLP, LLMs, agentic AI, travel intelligence, and geospatial AI.