Multimodal Adaptation
Developing efficient ways to adapt multimodal models to new domains, tasks, and operating constraints.
AI researcher · Multimodal intelligence
Intelligence isn’t static. It should know how to adapt as the world evolves.
I am an Applied Research Scientist at Openstream AI, developing multimodal and agentic systems that learn from diverse signals and remain useful beyond their training distribution.
Research focus
Building intelligence that can adapt, reason, and earn trust.
Multimodal Reasoning
Helping models connect visual evidence, language, and structured knowledge to solve complex reasoning tasks.
Trustworthy AI Agents
Building agents that learn from interaction, adapt over time, and make verifiable decisions in real workflows.
Recent research
All publications →2026
OntoForm: Ontology-Grounded Semantic Data Generation for Enterprise Document Understanding
EMNLP 2026 · Industry Track
A semantic-first framework for generating privacy-preserving enterprise form packets with ontology-grounded extraction, question answering, and validation signals.