Job
In this role, you select models for speech and signal processing, optimize their inference, and implement powerful processing pipelines to market maturity using PyTorch, LibTorch, and NVIDIA Triton.
This listing has been machine translated. Show original
Job technologies
Your role in the team
- Select suitable models, adapt and fine-tune them for signal and speech processing, including CNN and Transformer architectures, state-space models, and end-to-end ASR.
- Optimize inference through quantization, pruning, kernel fusion, batching strategies, mixed precision, TorchScript/torch.compile, as well as export to ONNX/TensorRT.
- Implement high-performance inference pipelines with PyTorch and LibTorch (PyTorch C++).
- Deploy and operate models on the NVIDIA Triton Inference Server, including Model Repository, Ensembles, Business Logic Scripting, Dynamic Batching, and custom Backends.
- Build the processing pipelines around the models, from signal preprocessing and feature extraction to streaming pipelines and pre- and post-processing.
- Translate prototypes to market maturity, take care of benchmarking, profiling, test coverage, reproducibility, monitoring, and model versioning, work closely with system architecture and product management, and evaluate new methods and tools regarding their practical suitability for our performance goals.
Our expectations of you
Education
- A completed degree in Computer Science, Electrical Engineering, Physics, Mathematics, or a comparable qualification, or alternatively, equivalent practical experience.
Qualifications
- Solid knowledge in digital signal processing (FFT, filter design, modulation, spectrogram representations) or in speech processing (ASR, speaker embeddings, speech enhancement).
- Excellent German and English skills complete your profile.
Experience
- Several years of experience with PyTorch beyond training: You know where models become slow in production and how to troubleshoot performance bottlenecks.
- Solid C++ skills from C++17 onwards, experience with the PyTorch C++ API, and practical knowledge in working with the Triton Inference Server you have already gained.
- In the past, you have gained experience with profiling and GPU optimization (CUDA fundamentals) and have a confident handling of Linux, Git, Docker, and CI/CD.
Benefits
Health, fitness & fun
Topics you will work on
Job locations
About your employer
PLATH GmbH & Co. KG, based in Hamburg, is a leading system integrator specializing in data-driven crisis detection. With a wide portfolio of signal processing solutions, the company serves the security and defense sector.
- Established Company
- Onsite, Hybrid
- Industry, Production
More jobs
These jobs might also interest you
The role As a Forward Deployed AI Engineer, you go directly into our specialist departments, understand the processes, and develop production-ready AI and automation solutions that will fundamentally change the way the entire GGW Group operates. You do...
sino AG, based in downtown Düsseldorf, has been an innovative financial service provider for over 25 years. We specialize in the execution of securities transactions for the most active and demanding traders in Germany. With currently 26 employees, we...
As a Senior Software Engineer (m/f/d), you will take on a central technical role in challenging client projects and develop sustainable, scalable software solutions in complex enterprise environments. In your role, you will primarily be responsible for:...