Prompt Order Analysis in VLMs
Investigating how prompt ordering affects vision-language model performance across tasks and providers
Publications and projects
Publications
Proper Scoring Rules for Agentic Uncertainty Quantification
Accepted at NeurIPS 2026. Also accepted at the CTB and FAGEN Workshops at ICML 2026.
SELFDOUBT: Uncertainty Quantification for Reasoning LLMs via the Hedge-to-Verify Ratio
Accepted as a poster at the FAGEN Workshop at ICML 2026
Don't Blink: Evidence Collapse during Multimodal Reasoning
Accepted at the Actionable Interpretability and Scientific Understanding of Foundation Models Workshops at COLM 2026
Confidently Wrong, Silently So: Auditing Undetectable Failures of a Deployed On-Device Language Model
Accepted at the NeurIPS 2026 Workshop on On-Device Intelligence (ODI 2026).
Selected Projects
3-Tiered Data Extraction from Images
Hybrid OCR and LLM system for intelligent information extraction
- Code: GitHub Repository
- Write-up: Medium Article
Multimodal Image Search with GPT-4 Vision
Combining visual and textual analysis for improved image retrieval
- Code: GitHub Repository
- Write-up: Medium Article
Image Aesthetics Quantification
CLIP-based aesthetics scoring system for hotel media ranking with domain adaptation
- Code: GitHub Repository
- Write-up: Medium Article