Senior Research Scientist | Generative AI | Computer Vision | Deep Learning | Digital Twins | Industrial AI

I am a Senior Research Scientist specializing in Generative AI, Computer Vision, and Multimodal Learning for civil infrastructure and large-scale industrial systems.

I hold a Ph.D. in Civil Engineering with a strong emphasis on applied computer vision and machine learning, and I have worked across academia, industry, and applied R&D to build scalable, production-ready AI systems.

Current and recent roles include:

• Senior Research Scientist, Augrade Inc., USA — Generative AI & Computer Vision
• Research Scientist, Siemens AI, USA — Industrial Computer Vision & Digital Twins
• Senior Research Associate, University of Southern California (Dept. of Civil Engineering)
• Journal Reviewer — SAGE Structural Health Monitoring, Elsevier Automation in Construction, Elsevier Computer Physics Communications, Springer Nonlinear Dynamics, and Springer The Journal of Supercomputing

My research lies at the intersection of Generative AI, Computer Vision, Deep Learning, and Image Processing, with a strong emphasis on real-world deployment and data-centric AI.

Core focus areas include:

• Generative AI

  • Synthetic data generation for vision tasks
  • GANs, DDPMs, Latent Diffusion Models (LDMs)
  • Unpaired image-to-image translation
  • Semantic mask synthesis and augmentation

• Computer Vision & Deep Learning

  • Crack detection, segmentation, and change detection
  • Semantic segmentation, object detection, defect classification
  • CNNs, Transformers, RNNs, and hybrid architectures
  • Classical + deep learning fusion pipelines

• Synthetic Dataset Generation

  • Large-scale synthetic pipelines for:
    • Structural cracks (concrete, pavement, steel)
    • AEC drawings (architectural, structural, MEP)
  • Algorithmic crack generation (~1.5s per sample)
  • Data diversity beyond physics-based CG and FEM approaches

• Digital Twins & Industrial AI

  • 2D-to-3D CAD model generation
  • Digital twin synchronization using:
    • Point clouds
    • RGB and RGB-D images
    • Multimodal sensor data
  • Graph-based representations and hierarchical assembly graphs
  • GNN-based parametrization and reconfiguration

• Multimodal Data

  • RGB, depth, RGB-D
  • 3D point clouds
  • Image–geometry–graph fusion

• Concrete, pavement, and steel infrastructure
• Structural and mechanical systems
• AEC, industrial facilities, and large-scale assets
• Vision-based inspection and monitoring systems

My work is driven by:

• Data-centric AI
• Algorithmic rigor
• Scalable and reproducible pipelines
• Bridging academic research with industrial deployment

I focus on building systems that actually work in practice—not just benchmarks.

This GitHub profile contains:

• End-to-end synthetic data generation pipelines
• Computer vision tools for structural inspection
• Point cloud annotation and review tools
• Multimodal learning experiments
• Reproducible research-grade codebases

Most repositories emphasize:

• Clean architecture
• Reproducibility
• Performance-aware design
• Practical usability

I am open to:

• Research collaborations
• Industrial R&D projects
• Synthetic data and Generative AI projects
• Open-source collaborations

Best ways to reach me:

• LinkedIn (professional inquiries)
• GitHub Issues / Discussions (technical discussions)

This profile reflects ongoing research and applied development in Generative AI, Computer Vision, and Digital Twin technologies for civil and industrial AI systems.