Deep Learning Performance Engineer
- Shanghai, China
Candidate is responsible to develop and optimize machine learning product for generic purpose or common use in a certain domain specific field. This job is responsible to bring the best accuracy/performance as well as easy-of-use to customer and embed you in ambitious and diverse team and influences Intel's AI platform and strategy. In this position, you will have chance to collaborate with diversified teams on full stack machine learning optimization and explore the most advanced Intel hardware platforms.
The job scope may include:
- Drive and deliver peak performance solutions to community benchmarks like MLPerf.
- Deliver the benefits of Intel's latest CPU/GPU platform and ingenious software solution to deep learning community and Intel's customers.
- Optimize the performance of influential models from academic and industry research.
- Participate in product concept readiness proving by implementing prototype.
- Co-work with diversified teams in design review and product implementation.
- Work closely with colleagues or customers cross organization or cross GEO to collect requirement, solve technical issue or share knowledge.
The candidate should possess at least a Master degree of Computer Science with at least 3 years of experience (or related BS/MS degree in Computer field) or a bachelor degree with at least 5 years of experience.
The candidate should have following qualifications:
Language and and fundamental skill :
candidate should be solid on at least one of the following language and familiar with at least two
• Assembly language
It will be a plus to be familiar with the knowledge of below domains
• Computer architecture
• OS and language runtime environment
- High Performance Computing
• 2 year+ experience in data analysis, in image/video, natural language, statistic data or other field
• Experience in deep learning frameworks such as Pytoch, TensorFlow, , MXNet etc...
o Have improved performance for one of the frameworks is a plus
o Have implemented new model on one of the frameworks is a plus
o Have ported a model from one framework to another is a plus
o Have used existing model to improve state-of-art of empirical problem is a plus
• Candidate should be familiar with deep learning algorithms and models.
- Excellent performance analysis and tuning skills
- Proven track record of deliver result in time on short release cycle
- Excellent written and oral communication skills and be able to clearly communicate technical details and concepts
- Proven track record of analyzing and solving complex technical problems
- Excellent development and debugging skills
- Good proven record for cluster infrastructure management with High Performance computing(HPC), like load balance , high availability service, and scalability
Inside this Business Group
Intel Architecture, Graphics, and Software (IAGS) brings Intel's technical strategy to life. We have embraced the new reality of competing at a product and solution level-not just a transistor one. We take pride in reshaping the status quo and thinking exponentially to achieve what's never been done before. We've also built a culture of continuous learning and persistent leadership that provides opportunities to practice until perfection and filter ambitious ideas into execution.
CN JR0149525 Shanghai
Back to top