About Shakeel Mohammed
Simulation Engineer and ML Developer : FEM · CFD · AI Surrogates
I help engineers get faster answers from their simulations by combining FEM, CFD, and machine learning in one workflow.
What I do:
- FEM & Structural Simulation: Structural, thermal, and vibration analysis using Ansys, FEniCSx, and CalculiX. Including experimental validation and digital twin integration.
- CFD & Flow Analysis: Flow simulations with OpenFOAM — from meshing to post-processing.
- Machine Learning
- Surrogate models that replace slow FEM/CFD runs
- PINNs — physics-based neural networks that work with little or no data
- Neural operators (FNO, DeepONet) for fast multi-case predictions
- Inverse problems to find material properties from measurements
- Image analysis : microstructure segmentation and defect detection
- Time-series prediction : vibration and sensor data with LSTM
- Deployment: PyTorch → ONNX → C++. HPC-ready
Typical projects:
- Too many simulation runs? → I build a surrogate that replaces them
- Unknown material parameters? → I set up an inverse PINN to identify them
- Microstructure images to analyze? → I train a segmentation model
- Simulation data sitting unused? → I turn it into a prediction tool
Stack: Python · PyTorch · Julia · Ansys · FEniCSx · OpenFOAM · ONNX · C++ · MPI · OpenCV · Fortran
Strong in the overlap between physics and AI: a combination most engineers and most ML developers don't cover alone.
English
Native or bilingual
German
Conversational
Remote only
Primarily works remotely
Experience
- Spektra gmbh dresdenResearch AssociateRESEARCHNovember 2024 - May 2025 (6 months)Dresden, Germany- Developed a CAD-to-FEM simulation workflow to analyze unintended transverse vibrations in the SPEKTRA SE-21 electrodynamic shaker using PrePoMax (pre/post) and CalculiX (solver), including modal and harmonic response analyses (u₁/u₂, u₃/u₂).- Performed systematic tolerance and imperfection studies (local stiffness changes, assembly deviations, magnetic-force imbalance, gravity/orientation effects) and automated evaluation/sensitivity ranking using Python.- Validated simulation findings against experimental/geometry inputs by comparing with laser vibration measurement data and 3D-scan–based deviation data with Opencv and plotly, translating results into design/manufacturing insights.
- Fraunhofer IWUResearch assistantRESEARCHApril 2024 - September 2024 (5 months)Dresden, Germany- Automated FEM simulation studies and result extraction using Ansys Workbench + Python (PyAnsys) to generate consistent datasets and speed up parametric investigations.- Designed and trained Deep Learning models in PyTorch for prediction on simulation/measurement datasets, including preprocessing, training, and evaluation and convert it onnx format for c++ integration.- Built sdk of digital twin by integrating onnx model with C++ components.
Recommendations
Be the first to recommend Shakeel Mohammed
Help this freelancer shine by sharing your experience working together.
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Certifications
- PyTorch for Deep LearningDeeplearning.ai2026