|
My research lies at the intersection of Computer Vision (CV) and Natural Language Processing (NLP), with a strong focus on their application in the Architecture, Engineering, and Construction (AEC) industry. I am passionate about building intelligent systems that can understand, reason, and interact with our physical world.
My specific research interests and contributions include:
- Foundational & Multimodal Models: Exploring the capabilities of large-scale models (e.g., SAM, Grounding DINO) and developing privacy-centric, locally-deployed LLMs for specialized domains.
- Retrieval-Augmented Generation: Building robust, hierarchical RAG systems for domain-specific knowledge retrieval, achieving over 1.5x accuracy improvement in construction code verification.
- Domain Adaptation & Synthetic Data: Bridging the sim-to-real gap by leveraging synthetic data from BIM and digital twins, achieving a 3x improvement in Average Precision (AP) for on-site segmentation tasks.

