I’m a B.Tech + M.S. by Research student in Computer Science and Computational Linguistics at IIIT Hyderabad, working on NLP, vision-language models, and AI alignment with Precog and LTRC. In 2025 I was a visiting researcher at Maastricht University, applying ML to sustainable agriculture under the EU Horizon project AMBROSIA.
- NLP
- Computational Social Science
- Graphs
- Applied ML
- Multimodal Systems
- Cognitive Science
- Responsible AI
Selected work
Adapting Multilingual Models to Code-Mixed Tasks via Model Merging
CODS-IKDD 2025 · 2025
A model-merging approach for adapting multilingual language models to code-mixed NLP tasks.
P. Kodali, V. Shivkumar, S. Joshi, M. Choudhary, P. Kumaraguru. (2025). "Adapting Multilingual Models to Code-Mixed Tasks via Model Merging." CODS-IKDD 2025.
SceneGraMMi: Scene Graph-boosted Hybrid-fusion for Multi-Modal Misinformation Veracity Prediction
arXiv:2410.15517 · 2024
A scene-graph-boosted hybrid-fusion architecture for predicting the veracity of multi-modal misinformation.
S. Joshi*, S. Mavani*, et al. (2024). "SceneGraMMi: Scene Graph-boosted Hybrid-fusion for Multi-Modal Misinformation Veracity Prediction." arXiv:2410.15517.
Read paperSee all publications · See all projects
Updates
- May 2026
Joined Fujitsu Research of India as an AI Research Intern, working on memory management for self-evolving multi-agent systems.
- Jan 2026
Helped develop and teach a session on Computational Psycholinguistics for 50+ students at IIIT Hyderabad.
- May 2025
Started as a Visiting Researcher at DACS, Maastricht University, on the EU Horizon project AMBROSIA.
- 2025
Two papers accepted: CODS-IKDD 2025 on model merging for code-mixed tasks, and IJCNLP-AACL 2025 (SRW) on compound semantics.
- 2025
Served as Program Chair for the MT Summit 2025 in Geneva, Switzerland.
- Oct 2024
Posted SceneGraMMi to arXiv, a scene-graph-boosted approach to multi-modal misinformation detection.