@Diogenese
Just did a keyword search for Xilinx and saw your post. Didn't realise we working with them back in 2017.
As you say, hope we're still in their good books somewhere in the background.
Catalyst for the word search was this job ad I just saw from a couple of weeks ago with Bosch. Not sure if posted previously.
The experience / knowledge area examples pretty much encompasses a lot of Akida connections and not saying has to do with us but Xilinx was the only one I didn't realise we had some history with.
![]()
Research Engineer - Deployment of Machine Learning Algorithms for Embedded Systems (f/m/div.)
jobs.smartrecruiters.com
Research Engineer - Deployment of Machine Learning Algorithms for Embedded Systems (f/m/div.)
- Robert-Bosch-Campus 1, 71272 Renningen, Germany
- Full-time
- Legal Entity: Robert Bosch GmbH
Company Description
At Bosch, we shape the future by inventing high-quality technologies and services that spark enthusiasm and enrich people’s lives. Our promise to our associates is rock-solid: We grow together, we enjoy our work, and we inspire each other. Welcome to Bosch.
The Robert Bosch GmbH is looking forward to your application!
Job Description
Machine Learning and especially Deep Learning algorithms play an important role for future advanced driver assistance systems. To deploy these algorithms in a fast and reliable way on hardware components, automatic software solutions for HW/SW co-design are essential.
- Within our team you evaluate compiler frameworks for modern algorithms (e.g. Deep Neural Networks (DNNs)), and hardware platforms (e.g. dedicated accelerators, vector processors).
- You map embedded AI algorithms, especially DNNs, onto virtual prototypes, digital embedded systems and disruptive future hardware-technologies such as in-memory and neuromorphic computing devices.
- Furthermore, you are responsible for adapting open-source compiler frameworks (e.g., TVM, MLIR, etc.) onto target-specific ISA.
- You work with first-class academic partners in international projects and supervise students.
- Last but not least, you scout deployment solutions for disruptive technologies such as in-memory and neuromorphic computing regarding their practical applicability in Bosch products.
Qualifications
- Education: excellent Master/Diploma/PhD degree in Information Technology, Computer Science or a related field
- Personality: highly motivated and open for innovative and disruptive ideas
- Working Practice: independent and proactive team player with goal oriented working style and strong communication skills
- Experience and Knowledge:
- expert background in digital technology, computer systems and systems-on-chip (e.g., Xilinx, ARM, RISC-V) as well as programming experience (Python, C/C++) are required
- very good understanding of Machine Learning and Deep Neural Network frameworks is desired (e.g., PyTorch, Keras, TensorFlow),
- experience with compiler technologies
- Enthusiasm: keen interest in future technologies and trends with passion for innovation
- Languages: very good in English and German (written and spoken)
Additional Information
Help us Boosting Bosch with AI.
Apply today at the Bosch Center for Artificial Intelligence in just 3 minutes!
https://www.bosch-ai.com
www.bosch.com/research
Please submit all relevant documents (incl. curriculum vitae, motivation letter, certificates).
You want to work remotely or part-time - we offer great opportunities for mobile working as well as different part-time models. Feel free to contact us.
Need support during your application?
Kathrin Stipak (Human Resources)
+49(711)811-38015
Need further information about the job?
Falk Rehm (Functional Department)
+49 (711)811-11827
Job Location
Hi Fmf,
Yes there are certainly many intersections with Akida.
Interesting that they mention ARM and RISC-V in the same breath when we have just announced partnerships with both (RISC-V via SiFive). I'm still trying to untangle some crossed synapses in my wetware which link RISC-V and ARM, which seems to have been triggered by the proximate temporal overlap of the partnerships of BrainChip with ARM and SiFive, and the fact that SiFive's newest processor had been relegated to equivalence with a 2 year old ARM processor - before SiFive announced the Akida connexion - which made the ARM/Akida link inevitable in my mind.
Looking at a recent Bosch NN patent, you can see why they would be interested in getting some more current state-of-the-art input.
US10735660B2 Method and device for object identification: Priority 20180217
[0014] According to one specific embodiment, the input data signals are read in an RGB format or YUV format in the step of reading in, the input data signals being read in, in particular, with a resolution of 12 bits or 16 bits per color channel. The bit depth in this case relates to the number of color shades per color or shades of gray. The raw input data signals from the camera are generally provided in 12-16 bits. This range may be reduced to 8 bits by various transformations adapted over time by the camera control. The adaptation of the 16 bits to 8 bits in this case takes place dynamically within various control loops. In principle, however, a higher resolution of 12-16 bits is available for the artificial intelligence of the neural network.
It would probably be a few times faster and more energy-efficient if they had a 4-bit processor, with negligible loss of accuracy.