"Fall in love with some activity, and do it! Nobody ever figures out what life is all about, and it doesn't matter. Explore the world. Nearly everything is really interesting if you go into it deeply enough." - Richard Feynman
Udbhav Tripathi
Research Scientist @ Ansys CTO Office
AI4Science
I am a researcher working at the interface of Scientific Machine Learning (SciML), Geometric Deep Learning, and Scientific Foundation Models (SciFM). My work centers on developing scalable, physics-aware machine learning frameworks for scientific data — with a focus on 3D reconstruction, computational physics, and simulation-driven learning. I build end-to-end pipelines from large scale data generation, preprocessing, training large diffusion, transformer models and experimenting with other architectures for scientific applications. I'm passionate about integrating physics priors with AI, building large scale toolkits for AI4Science applications.
Technical Expertise
- Scientific Foundation Models (UPT, Poseidon etc.)
- Geometric Deep Learning (TRELLIS, LATTE3D, etc.)
- Transformer-based scientific and geometric models
- 3D geometry processing using PyVista, Open3D, and VTK
- 3D computer vision (ViT, NERF, Gaussian Splatting, etc.)
Email: udbhavtri9696@gmail.com
Location: Pune, India
Current Projects
- Scientific Foundation Models (SciFM): Building large-scale, physics-informed models for simulation-aware learning across fluid dynamics and structural mechanics.
- Physics-Guided 3D Reconstruction: Developing pipelines using NeRF, Gaussian Splatting, and mesh based techniques to reconstruct 3D geometries and vector fields from multiview renderings of complex geometries.
- SciML for PDEs and Dynamical Systems: Designing transformer and diffusion-based networks for modeling real-world time dependent scientific datasets using Foundation operators.
- Geometric Deep Learning: Generating 3D mesh from multiple image patches using tripalnes, pretrained multiview aware diffusion models.
Selected Publications
Physics-integrated deep learning for uncertainty quantification and reliability estimation of nonlinear dynamical systems
Probabilistic Engineering Mechanics, Vol. 72, Article 103419