Etched is revolutionizing AI computing by developing specialized AI chips that are hard-coded for specific model architectures. Their first product, Sohu, is designed specifically for transformers and delivers performance that significantly outpaces traditional GPU solutions like the B200. As a Software Engineer in the Burn-in Testing team, you'll play a crucial role in ensuring the reliability and stability of their high-performance Inference server hardware and software.
The position involves designing and implementing comprehensive automated test suites, executing burn-in tests, analyzing results, and collaborating with both hardware and software engineering teams. You'll work with various technologies including Python, Go, Linux, and containerization tools like Kubernetes. The role requires a strong background in software testing, system architecture, and performance analysis.
Etched offers an attractive benefits package including full medical coverage, housing subsidies, and daily meals. The company maintains a fully in-person work environment in Cupertino, emphasizing direct collaboration and the integration of engineering and research. They're looking for candidates who can thrive in a fast-paced environment and contribute to both engineering and research aspects of the company.
What makes this opportunity unique is Etched's commitment to the "Bitter Lesson" philosophy - believing that AI progress comes from increased computational power through specialized hardware. This position offers the chance to work at the intersection of AI and hardware optimization, contributing to the development of next-generation AI infrastructure that enables previously impossible applications like real-time video generation and advanced reasoning agents.