---
title: 'Inference Software Engineer at Etched'
canonical: 'https://feeny.ai/job/inference-software-engineer-etched-cupertino-mn7xst9k2h5e'
type: 'job'
last_seen: '2026-09-04'
---

# Inference Software Engineer at Etched

- **Company:** Etched
- **Location:** Cupertino, CA
- **Posted:** 2025-01-16
- **Last confirmed live:** 2026-09-04
- **Apply:** https://job-boards.greenhouse.io/etchedai/jobs/4626560007

## Job description

## About Etched

Etched is building AI chips that are hard-coded for individual model architectures. Our first product (Sohu) only supports transformers, but has an order of magnitude more throughput and lower 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.

## Key responsibilities

- Contribute to the architecture and design of the Sohu host software stack
- Implement high-performance, modular code across the complete Etched software stack, consisting of a mix of Rust, C++ and Python.
- Interface with firmware and drivers teams delivering highest-performance HW/SW stack.
- Work with AI model researchers and product-facing teams building out the Etched serving front-end.

Representative projects

- Build scheduling logic for handling continuous batching and real time inference
- Implement inference-time acceleration techniques such as speculative decoding, tree search, KV cache sharing, etc.
- Implement distributed networking primitives for efficient multi-server inference

You may be a good fit if you have

- Experience with C++ and Python
- Familiarity with transformer model architectures and inference serving stacks (vLLM, SGLang, etc.) or experience working in distributed inference/training environments
- Experience working cross-functionally in large software and hardware organizations

Strong candidates may also have

- Experience with Rust
- Familiarity with GPU kernels, the CUDA compilation stack and related tools, or other hardware accelerators
- Understanding of distributed systems, networking, and parallel programming

## Benefits

- Full medical, dental, and vision packages, with 100% of premium covered
- Housing subsidy of $2,000/month for those living within walking distance of the office
- Daily lunch and dinner in our office
- Relocation support for those moving to Cupertino

## How we’re different

Etched believes in the [Bitter Lesson](http://www.incompleteideas.net/IncIdeas/BitterLesson.html). 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 Cupertino, 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.

## About Etched

## Company Overview
- **One-liner**: Etched builds specialized ASIC chips (the "Sohu" chip) designed exclusively for transformer-based AI models, aiming to deliver dramatically faster and more cost-efficient inference than general-purpose GPUs like Nvidia's.
- **Entity Type**: Private (Series B / Venture Round)
- **Headquarters**: San Jose, California, United States (primary HQ); also has a major office in Cupertino, CA, and offices in Taiwan, Canada, the UK, and Bulgaria.
- **Founded**: 2022
- **Founders**: Gavin Uberti (CEO), Chris Zhu (Co-Founder), Robert Wachen (Co-Founder & President)

## Core Business
- **Primary industry/industries**: Computer Hardware Manufacturing, AI Infrastructure, Semiconductor Design
- **Target customers**: B2B; Enterprise AI labs, large language model (LLM) companies, cloud providers, and hyperscalers needing high-performance inference compute.
- **Mission or purpose statement**: "Building the hardware for superintelligence." The company's stated goal is to enable the next generation of AI models by solving the inference bottleneck.

## Products & Services
- **[Sohu ASIC Chip]**: A custom-designed "Transformer ASIC" that is hardwired to run transformer models (the architecture behind GPT, Claude, Gemini, etc.). It is designed to be orders of magnitude faster and more power-efficient than Nvidia's H100/B200 for inference.
- **[Frontier Inference Clusters]**: Full-stack systems—bundles of Sohu chips, custom-designed racks, and specialized software—sold as turnkey "frontier inference clusters" to customers.
- **[Software Stack]**: A custom compiler and runtime (including a TensorFlow TVM backend) to optimize model deployment on the Sohu hardware.

## Market Standing
- **Valuation/Market Cap**: $5 billion post-money valuation (as of a $500M round closed in December 2025, announced June 2026).
- **Key Metric**: $1 billion in contract orders for its product (as of June 2026). Annual revenue is reported at $40M (likely pre-revenue or early-stage contract bookings).
- **Total Funding**: $800 million to date.
    - Seed Round (May 2023): $5.4M led by Primary Venture Partners.
    - Series A (June 2024): $120M led by Positive Sum and Primary Venture Partners.
    - Series B / Venture Round (Dec 2025): $500M led by Stripes. Participants include VentureTech Alliance, Jane Street, Hudson River Trading, Two Sigma, Ribbit Capital.
    - Angel Investors: Andrej Karpathy, Geoffrey Hinton, Fei-Fei Li, Arthur Mensch (Mistral CEO), Scott Wu (Cognition CEO), Stanley Druckenmiller, Peter Thiel.
- **Notable Investors/Partners**: Stripes, Primary Venture Partners, Positive Sum, Jane Street, Hudson River Trading, Two Sigma, Ribbit Capital, VentureTech Alliance.
- **Growth Signals**
    - Headcount: 258 employees (+141% YoY growth).
    - Chip successfully manufactured by TSMC in early 2026.
    - $1B in customer orders already booked.
    - Growing team of ex-Apple, ex-Google, ex-Nvidia, ex-Intel talent.
    - 100 active job postings with high growth in hiring.

## Competitive Advantages
- **Architecture Moat**: Unlike GPUs that are general-purpose, the Sohu chip is a fixed-function ASIC that only runs transformer models. This allows it to be 10-100x more efficient on the dominant AI workload (inference) than any GPU.
- **First-Mover in a New Category**: If transformer models remain the dominant architecture (as most experts believe), Etched has a multi-year lead in designing silicon specifically for them.
- **Elite Team**: Founders are Harvard dropouts and Thiel Fellows. The engineering team includes top talent from Nvidia (including the architect of the HGX/DGX systems), Google (TPU team, DeepMind), Apple (iPhone silicon), Intel, and Broadcom. The team has deep experience shipping complex chips at scale.
- **Investor Signal**: Backing from top-tier AI researchers (Hinton, Karpathy, Li) and quantitative trading firms (Jane Street, Hudson River) signals deep technical and financial credibility.

## Strategic Focus
- **Scale Production**: Moving from TSMC tape-out to volume manufacturing of the Sohu chip.
- **Deliver Clusters**: Fulfilling the $1B order book by shipping complete inference clusters to customers.
- **Build the Team**: Aggressively hiring across hardware (ASIC, PCB, physical design), software (inference, firmware, ML research), and operations (supply chain, finance, legal) to scale from ~250 to potentially thousands of employees.
- **Achieve "Gigawatt Scale"**: The company's internal mantra ("every decision moves us toward Gigawatt scale") indicates a focus on building infrastructure that can support massive AI workloads.

## Why Work Here
- **High-Impact, High-Ambition Environment**: Described as "end-to-end owners" who "own company-critical outcomes." The culture is built for people who want to move fast and solve hard, unsolved problems in AI hardware.
- **Cutting-Edge Tech**: Work on the absolute frontier of chip design and AI inference. The company is building what could be the most important AI chip of the decade.
- **Exceptional Team**: Colleagues include veterans from Nvidia, Apple, Google, and DeepMind. Learning opportunities are immense.
- **Rapid Growth**: Headcount is doubling annually, offering significant career growth and ownership opportunities for early hires.
- **Culture Mantras**: "Production is the product," "Build the machine" (team-building is as important as the hardware).
- **Work Policy**: Primarily on-site in Cupertino/San Jose, CA. The company emphasizes in-person collaboration for hardware development.
- **Notable Perks**: Working on a mission to build "the hardware for superintelligence" with massive financial backing and a clear product-market fit signal ($1B in orders).

## Sources
1. [TechCrunch - Nvidia competitor Etched hits $5B valuation](https://techcrunch.com/2026/06/30/nvidia-competitor-etched-hits-5b-valuation-1b-in-sales-for-ai-chip/)
2. [Etched.com - Careers Page](https://www.etched.com/careers)
3. [LinkedIn - Etched Company Profile](https://www.linkedin.com/company/etched-ai)
4. [Greenhouse - Etched Job Board](http://job-boards.greenhouse.io/etchedai)

## Other roles at Etched

- [Finance](https://feeny.ai/job/finance-etched-cupertino-q5pr34qq2pq1) — Cupertino, CA
- [Head of PCB Layout & Design Infrastructure](https://feeny.ai/job/head-of-pcb-layout-design-infrastructure-etched-cupertino-64hrte8av7r4) — Cupertino, CA
- [Emulation Software Engineer](https://feeny.ai/job/emulation-software-engineer-etched-cupertino-ewh85c1azrp0) — Cupertino, CA
- [Core Operations](https://feeny.ai/job/core-operations-etched-cupertino-thxbnhr8sfka) — Cupertino, CA
- [Front-End CAD Engineer](https://feeny.ai/job/front-end-cad-engineer-etched-cupertino-xe530w183ytw) — Cupertino, CA
- [Power Optimization Engineer](https://feeny.ai/job/power-optimization-engineer-etched-cupertino-4r5snp027zqh) — Cupertino, CA
- [Front-End Power Engineer](https://feeny.ai/job/front-end-power-engineer-etched-cupertino-f4xpz8s8xhnt) — Cupertino, CA
- [PCB Hardware Validation Engineer](https://feeny.ai/job/pcb-hardware-validation-engineer-etched-cupertino-rk48ajsz38v9) — Cupertino, CA
- [Head of Legal](https://feeny.ai/job/head-of-legal-etched-cupertino-pm15zcke4y2t) — Cupertino, CA
- [IC Package Engineer](https://feeny.ai/job/ic-package-engineer-etched-cupertino-wsq9bdz60pa8) — Cupertino, CA