Embedded Software Engineer (Bare-Metal)

Ubitium GmbH; Germany (OH); On-site; USD 110,000 - 160,000; Full time

Job summary

Ubitium GmbH is seeking Embedded Software Engineers to build benchmarks, demos, and reference implementations that leverage our accelerator modes. You and your team will develop new software and port established open‑source libraries to our platform, accelerating them with accelerator kernels.

Focus areas include bare‑metal signal processing and ML algorithms, porting OpenBLAS and other libraries, low‑level debugging, and performance profiling.

Qualifications

• Strong C skills and bare-metal development experience.
• Experience with Python or similar scripting languages.
• Port existing code to new platforms; signal-processing and ML kernels.
• Familiarity with BLAS/LAPACK and accelerator kernels.
• Low-level debugging and profiling; RISC-V and CUDA/OpenCL experience.

Responsibilities

• Build benchmarks, demos, and reference implementations leveraging accelerator kernels.
• Develop software and port established open-source libraries to our platform.
• Identify missing kernels and implement new accelerator kernels.
• Work on bare-metal signal processing and ML algorithms.
• Collaborate to optimize performance and profiling.

Skills

C

Bare-metal development

Python

Porting code

Signal processing

BLAS/LAPACK

ML model inference

Debugging & profiling

RISC-V

CUDA/OpenCL

Education

Bachelor's degree or equivalent

Tools

OpenBLAS

Open-source libraries porting

Job description

We are looking for Embedded Software Engineers who will build benchmarks, demos, and reference implementations leveraging our special acceleration modes. You and your team will develop new software and port established open-source libraries to our platform, accelerating them by using existing or newly developed accelerator kernels. Focus areas include, but are not limited to: Bare-metal implementations of signal processing and machine learning algorithms utilising existing accelerator kernels; selection, porting and integration of open-source libraries (e.g. OpenBLAS); identification of missing kernels and implementation of new accelerator kernels.

Prospective team members require strong C skills and bare-metal development experience as well as experience with Python or similar scripting languages. Preferred experiences include porting existing code to new platforms; signal-processing (e.g., FFT, FIR); numerical / linear algebra libraries (BLAS / LAPACK); inference of machine learning models; low-level debugging and profiling skills; RISC-V experience; parallel programming models (CUDA, OpenCL, or similar).

Bachelor's degree (or equivalent experience) and 2+ years of hands-on experience in a relevant role.

Relocation required. Currently 4 days on-site, 1 day remote. Policy may change.