About Etched
Etched is building the world’s first AI inference system purpose-built for transformers - delivering over 10x higher performance and dramatically lower cost and latency than a B200. With Etched ASICs, you can build products that would be impossible with GPUs, like real-time video generation models and extremely deep & parallel chain-of-thought reasoning agents. 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.

Job Summary
We are hiring a Head of SoC Architecture to lead our architecture team driving Etched’s next-generation AI inference systems. This is not an incremental optimization role. You will own the end-to-end architectural direction for transformer-specific silicon, from high-level workload modeling and dataflow strategy to subsystem definition and SoC integration tradeoffs. You will set the standard for architectural rigor, performance modeling, and long-term technical differentiation.

You will partner closely with ASIC, software, and system leadership to ensure our chips and full systems deliver step-function improvements in performance. This is a foundational role shaping Etched’s long-term technical moat.

Key Responsibilities
Team Building and Technical Leadership

• Build and lead a world-class architecture team

• Establish architectural review processes and quality standards

• Mentor senior architects and cultivate deep bench strength in modeling and innovation

Architectural Vision and Strategy

• Define the multi-generation SoC architecture roadmap aligned to transformer evolution and emerging inference workloads

• Drive macro-architectural tradeoffs across compute, memory hierarchy, interconnect, I/O, and system topology

• Establish clear performance, power, and area targets grounded in workload-level modeling

Transformer-Optimized Compute Architecture

• Lead dataflow, scheduling, and memory architecture decisions for large-scale transformer inference

• Translate model characteristics into hardware primitives that maximize throughput and minimize latency

• Identify architectural innovations that unlock new classes of real-time and agentic workloads

Performance Modeling and Analysis

• Oversee cycle-accurate simulation, analytical modeling, and workload characterization

• Build frameworks that connect model-level metrics to silicon-level constraints

• Ensure architectural decisions are quantitatively justified rather than intuition-driven

SoC Integration and Cross-Functional Leadership

• Partner with RTL, verification, physical design, firmware, and system software teams to ensure architectural feasibility

• Own architectural specifications, interface definitions, and subsystem boundaries

• Resolve cross-functional tradeoffs across power delivery, floorplan constraints, packaging, and bring-up realities

You may be a good fit if you have

• 12+ years of experience in computer architecture and ASIC design, with demonstrated ownership of complex SoC architectures

• Proven leadership defining high-performance compute or accelerator architectures

• Deep understanding of memory hierarchies, interconnects, compute pipelines, and system bottlenecks

• Experience driving architectural decisions through tapeout and silicon bring-up

• Strong background in performance modeling and workload-driven design

• Experience with AI/ML accelerator architectures or large-scale compute systems

• Strong intuition for power, area, and frequency tradeoffs at advanced process nodes

• Familiarity with cycle-accurate simulators and architectural modeling tools

• Ability to translate high-level workload requirements into microarchitectural specifications

Benefits

• Medical, dental, and vision packages with generous premium coverage

  • $500 per month credit for waiving medical benefits

• Housing subsidy of $2k 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

How we’re different

Etched believes in the Bitter Lesson. We think most of the progress in the AI field has come from using more FLOPs to train and run models, and the best way to get more FLOPs is to build model-specific hardware. Larger and larger training runs encourage companies to consolidate around fewer model architectures, which creates a market for single-model ASICs.

We are a fully in-person team in San Jose and Taipei, 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 as needed.