Job
- Level
- Experienced
- Location
- Darmstadt
- Working Model
- Onsite
- Job Field
- Software, Data
- Employment Type
- Full Time
- Contract Type
- Permanent employment
Job Summary
In this role, you will develop machine learning models to optimize and accelerate multiphysics simulations in Digital Twin environments, collaborating closely with physicists and engineers.
Job Technologies
Your role in the team
- This role sits at the intersection of ML engineering and computational physics, working within our Software Engineering & Digital Twin environments.
- You'll collaborate closely with optical engineers, laser and fusion physicists, simulation scientists, and systems engineers to embed data-driven intelligence directly into our Digital Twin architecture.
- You'll be reducing design risk, shortening development timelines, and unlocking system-level optimization that purely physics-based approaches can't achieve at the speed we need.
- You will design, build, and deploy machine learning models that enhance, accelerate, and augment our multiphysics simulation environments across our high-energy laser systems, target injection and tracking system, target manufacturing system, and fusion chamber.
- This will enable faster design cycles, predictive system optimization, and real-time decision support across our laser, targetry, and fusion programs.
- Surrogate & reduced-order modeling: Design and deploy surrogate and reduced-order models (ROMs) that replace or accelerate high-fidelity multiphysics simulations in the Digital Twin environment.
- Physics-informed ML: Develop physics-informed machine learning (PIML) and physics-informed neural networks (PINNs) that embed physical constraints (Maxwell's equations, thermodynamics, fluid dynamics) directly into model architectures.
- ML pipelines: Build and maintain ML pipelines for training, validation, uncertainty quantification (UQ), and continuous model refinement against experimental and simulation data.
- Active learning & Bayesian optimization: Implement active learning and Bayesian optimization workflows to intelligently guide design space exploration and reduce costly simulation runs.
- Digital Twin integration: Integrate trained ML models into the broader Digital Twin framework, interfacing with HPC simulation outputs (COMSOL, ANSYS, custom solvers) and real-time sensor data.
- Anomaly detection: Develop anomaly detection and predictive diagnostics models to monitor system health and identify off-nominal behavior in laser subsystems.
- Autonomous optimization: Apply reinforcement learning and Bayesian control approaches to support autonomous or semi-autonomous optimization of laser operating parameters.
- Cross-functional collaboration: Collaborate with Digital Twin architects, systems engineers, and optical simulation scientists to ensure ML models meet fidelity, latency, and uncertainty requirements.
- Engineering rigor: Establish best practices for model versioning, reproducibility, testing, and documentation in a fast-moving research environment.
This text has been machine translated. Show original
Our expectations of you
Qualifications
- Must-Haves
- Master's or PhD in Machine Learning, Computational Physics, Applied Mathematics, Data Science, Computer Science, or a closely related field.
- Fluency working with PDE-based simulation outputs, time-series sensor data, and high-dimensional parameter spaces typical of multiphysics environments.
- Ability to communicate model behavior, confidence intervals, and limitations clearly to physicists, engineers, and non-ML specialists.
- Strong cross-functional team skills; comfort working in an interdisciplinary environment spanning physics, engineering, and software.
- Nice-to-Haves
- Background in laser physics, plasma physics, high-energy-density science, or related complex physics domains.
- Vertrautheit mit interdisziplinären Design-Optimierungs-Workflows (MDO), einschließlich Design of Experiments (DoE), Sensitivitätsanalyse und Unsicherheitsausbreitung.
- Vertrautheit mit Monte Carlo-Methoden und Frameworks zur Quantifizierung statistischer Unsicherheiten.
- Interest in fusion energy, advanced laser systems, or high-energy-density physics.
Experience
- Proven experience building, training, and deploying ML models for complex physical systems; strong command of deep learning (PyTorch, TensorFlow/JAX), probabilistic models, and uncertainty quantification.
- Expert-level Python; proficiency in C++ or Fortran a plus; experience with HPC environments (batch schedulers, MPI/OpenMP parallelization).
- Hands-on experience with Gaussian processes, neural network surrogates, reduced-order models, or equivalent metamodeling techniques.
- Experience building robust ML pipelines for scientific data - including preprocessing, feature engineering, model validation, and deployment.
- Experience with physics-informed neural networks (PINNs) or neural operators (DeepONet, FNO) applied to physical systems.
- Experience with digital twin platforms and live integration of ML models with simulation environments (e.g., NVIDIA Omniverse, Siemens Xcelerator, ANSYS Twin Builder).
- Experience applying reinforcement learning to physical system control or optimization.
- Experience with MLOps tooling (MLflow, Weights & Biases, DVC) in a scientific computing context.
This text has been machine translated. Show original
What we offer
- Compensation offered will be determined by factors such as location, level, job-related knowledge, skills, and experience.
- Certain roles may be eligible for incentive compensation, equity, benefits.
This text has been machine translated. Show original
Topics You Will Work On
Job Locations
About Your Employer
Focused Energy
Focused Energy GmbH is a German-American company with offices in Darmstadt and Berlin, specializing in the development of technologies for laser-driven fusion. The company aims to make fusion energy available on an industrial scale, combining scientific and technical expertise from Germany and the USA.
Description
- Company Type
- Startup
- Working Model
- Onsite
- Industry
- Power Sector, Economy
Machine Learning Engineer
- Location
- Darmstadt
- Working Model
- Onsite
- Diversity
- Open for all genders
- English Only
- English only required
