About the job
About Etched
Etched is pioneering an AI inference system uniquely designed for transformers, achieving over 10 times higher performance while drastically reducing costs and latency compared to traditional systems like the B200. Our innovative ASICs enable groundbreaking products, including real-time video generation models and highly complex reasoning agents previously unattainable with GPUs. With substantial backing from top-tier investors and a team of elite engineers, Etched is transforming the infrastructure landscape for one of the fastest-growing industries in history.
Job Summary
As an Infrastructure Software Engineer, you will play a crucial role in developing model-specific ASICs by creating tailored infrastructure and toolchains that enhance the speed, reliability, and scalability of our development processes, from simulation to silicon, and from manufacturing to testing across various configurations. Our approach integrates software development best practices, ensuring the same level of rigor and quality in our infrastructure as we apply to our products.
Your responsibilities will include leading the development and implementation of next-generation infrastructure tools that empower our ASIC, software, and platform engineers to enhance productivity, reliability, and precision in manufacturing while pushing AI performance limits. This entails building and optimizing our hybrid high-performance compute (HPC) clusters for extensive CI, EDA workflows, emulation, and hardware-aware job execution.
Additionally, you will design and implement a cutting-edge observability stack featuring LLM integration, focusing on streaming health and performance telemetry, log aggregation, distributed tracing, insight generation, synthetic testing, and intelligent alerting across CI pipelines, simulation clusters, and service endpoints.
This position requires a robust software engineering mindset, an eye for quality, and a profound understanding of systems, it's not merely about writing scripts, but about crafting precise, repeatable infrastructure solutions with intention.
