Internal Acceleration, Engineer (ASIC)

JOB SUMMARY
Our Internal Acceleration team’s mandate is to build agents to accelerate the production of frontier tokens.
You’ll work directly with our founders and embed with the teams building the product (kernels, ASIC, hardware, production, and special projects) to build new agentic harnesses and operating systems to transform the way hardware is built in an AI native world. Your work will vary widely, and your success is measured by how much faster Etched moves, not by the lines of code you write.
This role is for engineers who thrive on ambiguity, have real product sense, and want to own problems end to end. It’s a whiteboard for future founders. The problems we solve are time-sensitive and business-critical. This role will not be a 9-5, and the ideal candidate is excited about that.
YOU MAY BE A GOOD FIT IF YOU
- Have at least 1-5+ years of experience building and shipping software end to end, ideally at a fast-moving startup
- Have strong full-stack skills. We care less about whether you write every line yourself and more about whether what you ship works
- Have built LLM agents or AI tooling that real users depend on, which means more than calling an API: context engineering, tool integration, evals, and failure analysis
- Use AI to ramp fast on unfamiliar domains. Agentic coding tools and frontier models are how you work, not an add-on
- Have product sense, design taste, and extreme attention to detail
- Have extreme agency: you find the real problem, not just the stated one
KEY RESPONSIBILITIES
- Embed with kernels, ASIC, hardware, and production teams to find the highest-leverage bottlenecks, and own each one from first idea to a tool engineers rely on every day.
- Build and deploy LLM agents wired into our engineering workflows via tool-calling and MCP.
- Ship full-stack internal products from zero to one: the interfaces, data pipelines, and infrastructure that define how Etched builds.
- Measure what works. Design evals and track adoption so we double down on the tools that matter and kill the ones that don’t.
- Take on special projects directly with the founders, surfacing problems across the company and doing whatever it takes to hit critical milestones.
STRONG CANDIDATES MAY ALSO HAVE
- Exposure to chip development, kernels, hardware, or manufacturing.
- Infrastructure, data engineering, or DevOps experience
- Fine-tuning, RAG over proprietary technical data, or multi-agent orchestration
- Startup experience, especially through hypergrowth
BENEFITS
- Medical, dental, and vision packages with generous premium coverage
- $500 per month credit for waiving medical benefits
- Housing subsidy of $2,500 per month for those living within walking distance of the office
- Relocation support for those moving to San Jose (Santana Row)
- Various wellness benefits covering fitness, mental health, and more
- Daily lunch and dinner in our office
- Unlimited compute budget subject to ROI justification
BASE COMPENSATION RANGE
- $150K – $400K+
- Plus Significant Equity
HOW WE’RE DIFFERENT
Etched believes in the Bitter Lesson. We are the first inference-focused frontier AI system, betting early on transformer and transformer-like architectures and on increasing model sizes. Our addressable market is the entirety of inference, unlike many of our competitors.
We are a fully in-person team in San Jose (Santana Row), and greatly value engineering skills. We do not have boundaries between engineering and research, and we expect all of our technical staff to contribute to both and work across disciplines as needed.
ABOUT ETCHED
Etched is building hardware for frontier intelligence. We co-design chips, racks, software, and manufacturing to deliver best-in-class throughput and latency across both prefill and decode workloads. Our first products are heavily focused on inference. Backed by hundreds of millions from top-tier investors and staffed by leading engineers, Etched is redefining the infrastructure layer for the fastest growing industry in history.

Employment type
FullTime

Work arrangement
No