The Role
We are seeking an architect to help ensure that data science applications and the most important learning frameworks leveraging the Python ecosystem offer industry-leading performance when run on the POWER microprocessor.
Your Impact
As part of the team that designed the #1 & #2 fastest supercomputers in the world, you will be part of a high profile initiative to optimize performance across all layers of the Python AI software stack for Power. This work will include libraries, Python packages, and learning frameworks such as TensorFlow and PyTorch for Power. This project provides opportunities to innovate in software optimization as well as exploitation of bleeding edge hardware architectures in upcoming processors.
Description
This role will involve identifying the libraries, frameworks and packages in the Python ecosystem that are most critical to machine learning/deep learning engineers and data scientists. Industry-standard benchmarks as well as open-source data sets and customer workloads will be used to stress this code in order to identify opportunities across the stack for performance improvement. Code will also be run on competitive platforms to understand relative areas of strength and weakness. Deliverables will include code implementation and improvements, tuning recommendations and whitepapers as well as sales collateral, and providing performance feedback to open source communities. Insights gained will be fed to hardware design teams to improve next-generation hardware.
Key Qualifications
by via developer jobs - Stack Overflow
We are seeking an architect to help ensure that data science applications and the most important learning frameworks leveraging the Python ecosystem offer industry-leading performance when run on the POWER microprocessor.
Your Impact
As part of the team that designed the #1 & #2 fastest supercomputers in the world, you will be part of a high profile initiative to optimize performance across all layers of the Python AI software stack for Power. This work will include libraries, Python packages, and learning frameworks such as TensorFlow and PyTorch for Power. This project provides opportunities to innovate in software optimization as well as exploitation of bleeding edge hardware architectures in upcoming processors.
Description
This role will involve identifying the libraries, frameworks and packages in the Python ecosystem that are most critical to machine learning/deep learning engineers and data scientists. Industry-standard benchmarks as well as open-source data sets and customer workloads will be used to stress this code in order to identify opportunities across the stack for performance improvement. Code will also be run on competitive platforms to understand relative areas of strength and weakness. Deliverables will include code implementation and improvements, tuning recommendations and whitepapers as well as sales collateral, and providing performance feedback to open source communities. Insights gained will be fed to hardware design teams to improve next-generation hardware.
Key Qualifications
- Prior experience working with the Python community and CPython implementation.
- Deep understanding of underlying mathematical operations that drive ML and Neural Networks
- Experience with prototyping, extending the architecture and optimizing to leverage multi-threading, SIMD, or other accelerators such as GPUs or FPGAs
- Strong understanding of memory layout, micro-architecture implications and experience working with compilers and interpreters
- Hands-on experience with projects utilizing machine learning and deep learning techniques with at least one of the following - TensorFlow, PyTorch, Caffe.
- Strong skills in Python, C, C++.
- Previous experience contributing to open source projects.
- Ability to work in a team, network with people outside of the team and effectively communicate in written and verbal presentations is essential.
- Experience implementing or optimizing key Python packages such as NumPy, pandas, SciPy would be an asset.
*Required Professional and Technical Expertise
In this role, we require:
- 5+ years in Python programming
- Demonstrated experience implementing neural network or machine-learning algorithms
- Experience with the internals of least one of TensorFlow or PyTorch (performance optimizations, feature additions, bugfixes etc.)
- Knowledge of modern microprocessor design
- Experience using acceleration technologies for data science applications (GPU, FPGA, SIMD, custom ASIC etc.)
- Passion for continuous improvement in building knowledge base both technically and professionally
- Minimum BS OR MS degree in Computer Science, Computer Engineering or a related technical discipline or equivalent experience.
*Preferred Professional and Technical Expertise
- Experience with multiple frameworks
- Recognized status in a key open source community
- In-depth knowledge of AIs application in industry
by via developer jobs - Stack Overflow
No comments:
Post a Comment