
Surrogate Modeler
CFD Research
- Huntsville, AL
- Permanent
- Full-time
The Applied Surrogate Modeling team is a small group of multi-disciplinary engineers that focuses on the development and deployment of fast-running, data-driven surrogate models for mission-specific applications.Our team is responsible for: * Generating DoD system performance data using physics-based tools such as computational fluid dynamics (CFD) and finite element methods (FEM)
- Training surrogate models to approximate these physics-based processes
- Deploying these surrogate models into systems that require high inference rates (1+ kHz)
- Training data-driven surrogate models that approximate a wide variety of physics-based systems using the Sage software
- Exporting and deploying surrogate models into high-throughput applications such as 6DOF simulators, hardware/software-in-the-loop systems, and optimization workflows
- Demonstrating, validating, and verifying surrogate model accuracy
- Working with other engineers to guide and inform physics-based data collection using design of experiments techniques
- Supporting development of new surrogate modeling techniques where necessary
- Candidate must be a U.S. Citizen and meet eligibility to obtain/maintain a SECRET Clearance
- Master's in Aerospace or Mechanical Engineering, Computer Science, or similar
- Proficiency with Python and Linux-based operating systems
- Experience working with coupled, multi-physics simulations and data including: external aerodynamics, heat transfer, structural modeling, and aerothermoelasticity
- Familiarity with machine-learning and surrogate-modeling techniques such as: artificial neural networks (ANNs), Kriging, and Gaussian process regression
- Ph.D. in Aerospace or Mechanical Engineering, or similar discipline
- Proficiency with compiled languages, especially C++ and Fortran
- Experience with the PyTorch machine learning library
- Knowledge of formal gradient-based and gradient-free optimization techniques
- Experience with multi-disciplinary analysis and optimization (MDAO)
- Active SECRET clearance