Welcome!

I am Abhishek Srivastava, a Project Associate in the Department of Mechanical Engineering at the Indian Institute of Technology (IIT) Palakkad, with a Ph.D. in Mechanical Engineering from IIT Delhi (2026). My research lies at the intersection of Scientific Machine Learning, Computational Fluid Dynamics, and Multiphysics Modeling, where I develop interpretable, data-efficient, and uncertainty-aware computational frameworks for solving complex engineering problems.

My doctoral research focused on integrating high-fidelity multiphysics simulations, computational fluid dynamics (CFD), electrostatics, explainable artificial intelligence (XAI), surrogate modeling, and multi-objective optimization to investigate charged particle dynamics and optimize electrostatic particle collection systems. By combining first-principles physics with machine learning, my work demonstrated how data-driven models can accelerate engineering design while preserving physical interpretability and computational efficiency.

I have extensive experience in Computational Fluid Dynamics (CFD), Multiphysics Simulation (COMSOL Multiphysics), Scientific Computing with Python, Explainable AI (SHAP and Partial Dependence Plots), Gradient Boosting, Surrogate Modeling, Engineering Optimization, and Data-Driven Modeling. At IIT Palakkad, my current research focuses on Bayesian inverse problems and uncertainty quantification: I develop variational-autoencoder-based posterior approximation methods benchmarked against MCMC reference posteriors, analyse when amortized variational posteriors under- or over-estimate uncertainty using the Fisher Information Matrix, and study parameter identifiability and optimal sensor placement for inverse heat transfer with physics-informed surrogates (PINN/PIKAN). The goal is trustworthy, uncertainty-aware AI methodology for computational science and engineering.

Research Interests

• Scientific Machine Learning
• Bayesian Inference
• Uncertainty Quantification
• Physics-Informed Machine Learning
• Deep Generative Models
• Bayesian Inverse Problems
• Computational Fluid Dynamics (CFD)
• Multiphysics Modeling
• Surrogate Modeling
• Explainable Artificial Intelligence (XAI)

Research Vision

My long-term research vision is to bridge physics-based modeling, numerical simulations, machine learning, and probabilistic inference to develop next-generation computational frameworks for engineering applications. I am particularly interested in uncertainty-aware AI systems that combine physical principles with modern machine learning to enable reliable prediction, inverse modeling, digital twins, and intelligent decision-making for complex physical systems.

Collaboration

I enjoy working on interdisciplinary problems that combine engineering, applied mathematics, and artificial intelligence. I welcome collaborations with researchers working in Scientific Machine Learning, Computational Mechanics, Computational Physics, Bayesian Methods, Uncertainty Quantification, Physics-Informed Learning, Deep Generative Models, and AI for Science.

I am currently seeking postdoctoral opportunities in Scientific Machine Learning, Bayesian AI, Computational Engineering, and Data-Driven Modeling, where I can contribute to cutting-edge research while collaborating with multidisciplinary teams to solve impactful real-world engineering challenges.

Thank you for visiting my website. Please feel free to explore my publications, research projects, and curriculum vitae, or get in touch if you are interested in collaboration or research opportunities.