Works across embodied robotics and aerial perception with multimodal vision‑language‑action models, and applies explainable, knowledge‑graph methods to civil materials and infrastructure sustainability. Emphasis on scalable learning, real‑time control, and reliable open datasets.