Anantha Krishnan O.

PhD | CFD/FEA | Data Scientist | ML | AI & Predictive Modeling I help businesses, researchers, and engineering teams turn complex data and simulation results into practical machine-learning models, predictive tools, CFD/FEA workflows, and reliable data-driven solutions. My work covers the full data-science lifecycle: data cleaning and exploration, feature engineering, model development, validation, optimization, visualization, and deployment-ready implementation. I primarily work with Python, PyTorch, Scikit-learn, Pandas, NumPy, SciPy, MATLAB, and C++. I hold a PhD in Mathematical Analysis, Modelling and Applications from SISSA, with research focused on mathematical modelling, computational engineering, scientific computing, finite-element methods, reduced-order modelling, and data-driven methods. This gives me a strong foundation for complex datasets, numerical systems, CFD/FEA simulation data, time series, optimization, and scientific or engineering applications. Selected experience and results: • PhD research: Built ML surrogates using Gaussian-process regression, neural networks, and autoencoders. An eight-mode POD-GPR model achieved 896× mean online speedup over the full-order solver with 0.109 mean relative L2 displacement error on 100 test cases in a six-parameter static benchmark. • AEE INTEC: Developed and integrated solar-thermal models in the open-source SunPeek D-CAT framework, including operational-data validation, sensor mapping, automated tests, and diagnostic plots. • Industrial software: Developed CFD/FEA workflows, multi-objective optimization algorithms, and C++ implementations for engineering simulation and additive manufacturing. • TU Berlin: Combined compressible CFD, simulation time series, signal processing, and MATLAB system identification to build low-order acoustic models. What I can help you with: • Machine Learning: Regression, classification, clustering, forecasting, anomaly detection, model selection, hyperparameter optimization, and predictive modelling • CFD/FEA & Engineering Simulation: Computational fluid dynamics, finite-element analysis, structural/thermal simulation, simulation-data analysis, model verification, and parametric studies • Data Science & Analytics: Data cleaning, preprocessing, EDA, statistical analysis, visualization, feature engineering, and insight generation • Deep Learning: Neural networks, PyTorch models, autoencoders, representation learning, and custom architectures • Time-Series & Sensor Data: Signal processing, operational-data analysis, forecasting, system identification, anomaly detection, and predictive modelling • Scientific Machine Learning: Physics-based and data-driven modelling, surrogate models, reduced-order models, and ML acceleration of computational simulations • Predictive & Surrogate Modeling: Fast approximations for expensive simulations, parameter prediction, response approximation, and design-space exploration • Optimization: Parameter optimization, sensitivity analysis, multi-objective optimization, and computational experiment design • Scientific Python Development: Numerical algorithms, automated analysis pipelines, custom modelling tools, data-processing workflows, and reproducible research software • Model Validation: Error analysis, cross-validation, benchmarking, statistical evaluation, model comparison, and performance optimization Technical Stack Python, PyTorch, Scikit-learn, Pandas, NumPy, SciPy, Matplotlib, Jupyter, MATLAB, C++, Git, SQL, FEniCS, RBniCS, EZyRB, OpenFOAM, ANSYS Fluent/CFX, and engineering CAE tools. My experience spans machine learning, computational modelling, CFD/FEA, engineering simulation, fluid dynamics, structural mechanics, thermal systems, sensor data, and scientific software development. This combination lets me assess not only whether a model performs well, but also whether it makes sense for the underlying physics, data, and real-world problem. I can support projects involving machine-learning model development, data analysis and visualization, predictive analytics, time-series forecasting, scientific and engineering datasets, neural networks and deep learning, surrogate modelling, simulation-data analysis, CFD/FEA workflows, optimization and parameter studies, Python data pipelines, statistical modelling and validation, and research-oriented machine learning. I focus on solutions that are accurate, explainable, reproducible, and useful in practice—not simply models with good metrics. If you have a dataset, prediction problem, CFD/FEA or simulation challenge, computational bottleneck, or an idea that needs to become a reliable machine-learning workflow, I can help take it from initial analysis through implementation and validation.