> Markdown version of [/jobs/ext/2380826-master-student-ai-modeling-for-wss-and-ocs-switching-systems](https://www.wearedevelopers.com/jobs/ext/2380826-master-student-ai-modeling-for-wss-and-ocs-switching-systems). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Master Student - AI Modeling for WSS and OCS Switching Systems

- **Company:** Huawei
- **Location:** München, Germany (Remote available)
- **Contract:** Permanent contract
- **Skills:** Artificial Intelligence, Artificial Neural Networks, Wavelength-Division Multiplexing, Tensorflow, Reinforcement Learning, Pytorch, Deep Learning
- **Published:** August 4, 2026
- **Apply:** https://www.xing.com/jobs/muenchen-master-student-ai-modeling-wss-ocs-switching-systems-157543655

## About the Role

* Enrolled in a Master's program in EE, Photonics, Applied Physics, Optical Engineering, CS, or related field.
 * Deep learning fundamentals; experience with PyTorch or TensorFlow
 * Familiarity with one or more: inverse design, physics-informed neural networks (PINNs), reinforcement learning
 * Understanding of metasurface operating principles (phase/amplitude/polarization control via subwavelength structures).
 * Knowledge of unit-cell design, lattice types, phase-gradient concepts, and fabrication constraints
 * Foundational knowledge of physical optics, liquid crystal optics, and LCOS device operation (both conventional and pixelated isotropic modes).
 * Understanding of WSS architecture and key metrics: insertion loss, crosstalk, bandwidth, PDL, port count.
 * Experience with LC device modeling (Frank-Oseen theory, director simulation).
 * Knowledge of DWDM systems and telecom optical component design
 * Prior work at the intersection of AI and photonics/optics
 * Ability to read technical literature independently.
 * Ability to work independently and collaboratively in a research-oriented environment

## Description

We are seeking a Master student to work on AI-driven modeling and optimization of metasurface-based Pixelated Isotropic LCOS (PI-LCOS) devices and Wavelength Selective Switch (WSS) systems. The project has two complementary tracks:

 1. PI-LCOS metasurface optimization - Use AI techniques to optimize metasurface shapes and topologies to reduce insertion loss and achieve low polarization-dependent loss (PDL) in metasurface-based PI-LCOS devices.
 2. WSS modeling & optimization - Develop AI-assisted models for WSS systems based on conventional LCOS technology, focusing on crosstalk suppression, insertion loss reduction, and system-level performance improvement.

 * Develop AI/ML surrogate models and inverse design frameworks for metasurface unit-cell optimization targeting PI-LCOS devices, with objectives including insertion loss minimization and PDL reduction.
 * Build and refine WSS system-level models (optical path, LCOS switching engine, port architecture) for conventional LCOS-based WSS, evaluating insertion loss, crosstalk, passband shape, and PDL
 * Investigate WSS crosstalk suppression strategies through pixel layout design, phase/amplitude apodization, and AI-driven optimization.
 * Perform electromagnetic simulations (FDTD, RCWA, FEM) of metasurface structures and validate AI predictions against full-wave results.
 * Optimize metasurface geometry parameters (pillar shape, lattice, orientation, material) to minimize insertion loss and PDL while meeting phase/amplitude/polarization requirements.
 * Apply AI/ML methods to the LCOS-based WSS design loop for multi-objective optimization across system-level performance metrics.


## Related Videos

- [Photonic Computing: Programming a New Class of AI Accelerators (incl. Live Coding)](https://www.wearedevelopers.com/videos/100196-photonic-computing-programming-a-new-class-of-ai-accelerators-incl-live-coding)
- [Getting Started with Machine Learning](https://www.wearedevelopers.com/videos/260-getting-started-with-machine-learning)
- [The pitfalls of Deep Learning - When Neural Networks are not the solution](https://www.wearedevelopers.com/videos/14-the-pitfalls-of-deep-learning-when-neural-networks-are-not-the-solution)
- [Machine learning in the browser with TensorFlowjs](https://www.wearedevelopers.com/videos/155-machine-learning-in-the-browser-with-tensorflowjs)
- [Geometric deep learning for drug discovery](https://www.wearedevelopers.com/videos/264-geometric-deep-learning-for-drug-discovery)
- [Schroedinger's cat: Thinking in- and outside the box of quantum mechanics](https://www.wearedevelopers.com/videos/216-schroedinger-s-cat-thinking-in-and-outside-the-box-of-quantum-mechanics)

## Related Articles

- [How We Built a Worry-Free System That Runs for 10+ Years – And What We’d Do Again](https://www.wearedevelopers.com/magazine/751-how-we-built-a-worry-free-system-that-runs-for-10-years-and-what-we-d-do-again)
- [Best Coding Boot Camps in Germany](https://www.wearedevelopers.com/magazine/237-best-coding-boot-camps-in-germany)
- [ I Gave a Video Editor More Autonomy Than a Trading Bot. On Purpose.](https://www.wearedevelopers.com/magazine/773-i-gave-a-video-editor-more-autonomy-than-a-trading-bot-on-purpose)
- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)
- [Dev Digest 134 - Where pixels sing?](https://www.wearedevelopers.com/magazine/477-dev-digest-134-where-pixels-sing)
- [What Are Large Language Models?](https://www.wearedevelopers.com/magazine/304-what-are-large-language-models)