OpenAI，这家知名的AI研究机构，正开启新的战略篇章，计划深度参与外部供应商的硬件设计流程，以定制符合其特定需求的AI硬件。据Data Centre Dynamics报道，OpenAI正在招聘软硬件协同设计工程师，这一举措旨在强化其在人工智能领域的硬件能力。
该职位的设立不仅要求工程师能够与OpenAI内部的机器学习工程师、内核工程师和编译器开发人员紧密协作，理解他们在高性能加速器方面的愿景和需求，比如在机器学习技术、算法、数值近似、编程表达性和编译器优化等方面的具体要求，还要求其具备与多个外部供应商合作的能力，共同推动AI硬件的性能提升和可编程性优化。
工程师将扮演桥梁的角色，协调第三方开发者设计最佳内核，并确保这些内核在OpenAI的编译器中得到充分支持。此外，他们还需要对不同硬件配置的关键内核进行性能估算，以确保硬件设计的效率和实用性。这一招聘信息来自于IT之家，表明OpenAI正逐步从软件层面深入到硬件层面，以实现更高效、更定制化的AI解决方案。
英语如下：
News Title: “OpenAI Enters Hardware Design Arena with External Partners for AI Future”
Keywords: OpenAI, AI hardware, supplier collaboration
News Content: OpenAI, the renowned AI research organization, is embarking on a new strategic chapter, planning to deeply engage in the hardware design process with external suppliers to create AI hardware tailored to its specific needs. As reported by Data Centre Dynamics, OpenAI is currently recruiting Co-Design Engineers for hardware and software, a move aimed at enhancing its hardware capabilities in the field of artificial intelligence.
The position necessitates engineers who can collaborate closely with OpenAI’s in-house machine learning engineers, kernel engineers, and compiler developers, understanding their vision and requirements for high-performance accelerators. This includes specific aspects such as machine learning techniques, algorithms, numerical approximations, programming expressiveness, and compiler optimizations. Additionally, the role requires the ability to work with multiple external suppliers to jointly advance AI hardware performance and programmability.
Engineers will act as a bridge, coordinating with third-party developers to design optimal kernels and ensuring their seamless integration within OpenAI’s compilers. They will also be responsible for estimating the performance of critical kernels across different hardware configurations, guaranteeing the efficiency and practicality of hardware designs. This job posting, sourced from IT Home, indicates OpenAI’s gradual progression from software to hardware realms in pursuit of more efficient and customized AI solutions.
【来源】https://www.ithome.com/0/754/534.htm
Views: 2