kmaehashi

Preferred Networks | Tokyo or Remote in Japan | Full-time | https://www.preferred.jp/en

Preferred Networks is an AI company based in Tokyo working across the stack, from AI chips and computing infrastructure to LLMs and products. You may already know us indirectly if you've used software we've built, such as Optuna or CuPy (or Chainer, back in the day).

We are designing in-house chips (MN-Core series: https://mn-core.com/) and training LLMs (PLaMo series: https://huggingface.co/pfnet), and our team is actively hiring for two roles related to these endeavors:

- MN-Core LLM Serving Engine Engineer: Build software infrastructure to serve LLMs using our upcoming inference accelerator, MN-Core L1000. [Apply here: https://open.talentio.com/r/1/c/preferred/pages/121580]

- LLM Inference Optimization Engineer: Improve the inference engine powering our API service and maintain PLaMo implementations in open source projects such as vLLM. [Apply here: https://open.talentio.com/r/1/c/preferred/pages/119173]

Both roles require relocation to Japan. We are happy to provide visa and relocation support.

Preferred Networks | Tokyo or Remote in Japan | Full-time | https://www.preferred.jp/en

Preferred Networks is an AI company based in Tokyo working across the stack, from AI chips and computing infrastructure to LLMs and products. You may already know us indirectly if you've used software we've built, such as Optuna or CuPy (or Chainer, back in the day).

We are designing in-house chips (MN-Core series: https://mn-core.com/) and training LLMs (PLaMo series: https://huggingface.co/pfnet), and our team is actively hiring for two roles related to these endeavors:

- MN-Core LLM Serving Engine Engineer: Build software infrastructure to serve LLMs using our upcoming inference accelerator, MN-Core L1000. [Apply here: https://open.talentio.com/r/1/c/preferred/pages/121580]

- LLM Inference Optimization Engineer: Improve the inference engine powering our API service and maintain PLaMo implementations in open source projects such as vLLM. [Apply here: https://open.talentio.com/r/1/c/preferred/pages/119173]

Both roles require relocation to Japan. We are happy to provide visa and relocation support.

Preferred Networks | Tokyo or Remote in Japan | Full-time | https://www.preferred.jp/en

Preferred Networks is an AI company based in Tokyo working across the stack, from AI chips and computing infrastructure to LLMs and products. You may already know us indirectly if you've used software we've built, such as Optuna or CuPy (or Chainer, back in the day).

We are designing in-house chips (MN-Core series: https://mn-core.com/) and training LLMs (PLaMo series: https://huggingface.co/pfnet), and our team is actively hiring for two roles related to these endeavors:

- MN-Core LLM Serving Engine Engineer: Build software infrastructure to serve LLMs using our upcoming inference accelerator, MN-Core L1000. [Apply here: https://open.talentio.com/r/1/c/preferred/pages/121580]

- LLM Inference Optimization Engineer: Improve the inference engine powering our API service and maintain PLaMo implementations in open source projects such as vLLM. [Apply here: https://open.talentio.com/r/1/c/preferred/pages/119173]

Both roles require relocation to Japan. We are happy to provide visa and relocation support.

CuPy isn't semi-abandoned as well, obviously :)

As a maintainer of CuPy and also as a user of several GPU-powered Python libraries, I empathize with the frustrations and difficulties here. Indeed, one thing CuPy values is to make the installation step as easy and universal as possible. We strive to keep the binary package footprint small (currently less than 100 MiB), keep dependencies to a minimum, support wide variety of platforms including Windows and aarch64, and do not require a specific CUDA Toolkit version.

If anyone reading this message has encountered a roadblock while installing CuPy, please reach out. I'd be glad to help you.

NumPy has a mechanism to dispatch execution to CuPy: https://numpy.org/neps/nep-0018-array-function-protocol.html

Just prepare the input on NumPy or CuPy, and then you can just feed it to NumPy APIs. NumPy functions will handle itself if the input is NumPy ndarray, or dispatch the execution to CuPy if the input is CuPy ndarray.

cuDF is a CuPy-based library providing drop-in replacement for Pandas: https://rapids.ai/

Actually all CUDA Toolkit libs are already available through the conda-forge channel: https://anaconda.org/conda-forge/cuda-cudart, https://anaconda.org/conda-forge/libcublas, etc.

CuPy tensors (or `ndarray`) provide the same semantics as NumPy. In-place operations are permitted.

For those interested in the NumPy/SciPy API coverage in CuPy, here is the comparison table:

https://docs.cupy.dev/en/latest/reference/comparison.html

An excellent example of Array API usage can be found in scikit-learn. Estimators written in NumPy are now operable on various backends courtesy of Array API compatible libraries such as CuPy and PyTorch.

https://scikit-learn.org/stable/modules/array_api.html

Disclosure: I'm a CuPy maintainer.