- Location
- San Jose, CA, United States
- Employment
- Full time
- Posted
- 19 November
About this role
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.
Job Summary
Etched builds custom chips and complete racks for AI inference. As deployments grow across racks and data centers, the network has to deliver the bandwidth, latency and reliability those systems depend on. Our Supercomputing team builds the software and infrastructure that make that possible.
We’re hiring networking engineers to help design, bring up and operate the networks connecting our inference systems. The work spans RDMA fabrics, network provisioning, configuration management and telemetry, with ownership from initial deployment through production. When performance drops or a link fails, this team builds the tools to understand what happened and get the system back to full capacity.
This is a hands-on role for someone who has worked with real data center networks and writes software to make them easier to deploy, operate and debug. You’ll work directly with the engineers building our hardware, firmware and system software, helping establish how Etched networks scale as we bring more systems online.
Key Responsibilities
Design and bring up data center networks connecting inference nodes within and across racks, including RDMA fabrics.
Build software to automate network provisioning, configuration, validation and recovery across deployments.
Develop safe processes for network changes, with testing, staged rollouts and rollback.
Build telemetry and diagnostics that expose network health, congestion and failures, and help engineers trace issues across hosts, NICs and switches.
Investigate bandwidth, latency and reliability issues under inference workloads, working across network configuration, Linux, drivers and firmware.
Develop automated qualification, burn-in and failure tests for NICs, switches and network configurations before deployment.
Partner with Node Systems, Fleet Software and deployment teams to integrate networking into system bring-up and ongoing fleet operations.
Work with hardware vendors to resolve issues and qualify new equipment and firmware.
You May Be a Good Fit If You Have
Experience building and operating production data center networks, with direct responsibility for deployment, troubleshooting and reliability.
Hands-on experience with RDMA networking, including RoCE, and an understanding of how fabric configuration and congestion affect performance.
Strong networking fundamentals, including Ethernet, IP routing, switching and network security.
Experience writing production software or automation for network provisioning, configuration management, monitoring or remediation.
Proficiency in a language such as Python, Go, C++ or Rust.
Strong Linux fundamentals and the ability to debug across hosts, NICs and switches using logs, counters, packet captures and other diagnostics.
Experience owning technical work from design through deployment and supporting it when things fail in production.
Comfort working across software, hardware and operations teams.
Strong Candidates May Also Have
Experience with WAN or backbone networks, including routing, traffic engineering and connectivity between data centers.
Experience operating CDN, cloud or other large-scale network infrastructure.
Experience with GPU, TPU or other accelerator clusters.
Familiarity with switch operating systems and network management interfaces such as SONiC, gNMI or OpenConfig.
Experience tuning congestion control and diagnosing packet loss, queueing and uneven traffic distribution in RDMA fabrics.
Experience building network telemetry pipelines and using them to detect failures or performance regressions.
Experience with network failure testing, failover and recovery at scale.
Representative Projects
Automate the provisioning and validation of a new rack’s network so it can enter service with a tested configuration.
Diagnose an RDMA performance regression by correlating host, NIC and switch telemetry under load.
Build a network rollout workflow that detects unhealthy changes and safely restores the previous configuration.
Develop tests that show how an inference cluster behaves when links, switches or nodes fail.
Benefits
Medical, dental, and vision packages with generous premium coverage
$500 per month credit for waiving medical benefits
Housing subsidy of $2.5k 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 + dinner in our office
Unlimited compute budget subject to ROI justification
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.
As published by Etched. Applications are handled on their site.
Skills this posting mentions
About Etched
Frontier Inference Clusters
All 116 openings at EtchedOne click, then it is written
Apply to Etched with a resume written for this role.
Queue MTS, Supercomputing Networking and I read the posting, rewrite your resume against it, draft the cover letter, and score the fit. Then you press send, or press one button and I fill in Etched’s form for you.
01Drop your resume
A PDF or a LinkedIn URL. About a minute, once.
02I rank the openings
Every weekday morning, the live catalog scored against your history. Up to 20 worth your time, not two hundred links.
03Each one is written up
Resume rewritten for the posting, a cover letter, a fit score. Press send, or let me fill in the form.
- 25 sent a week, free
- No card
- Nothing sent until you say so
More roles at Etched
See all- 19 November
- 18 November
- 16 November
- 12 November
- 12 November
- 12 November
Similar roles elsewhere
See more- 17 November
Electrical Hardware Engineer (multiple levels available)
Muon SpaceSan Jose, CA
$146k - $224k/yrMid levelEngineering - 17 November
- 13 November
Principal Design Verifiction Engineer
SambaNovaSan Jose, California
$180k - $225k/yrPrincipalEngineering - 13 November
Put this to work
Paste your career in once. Every application after that is written for you.
Drop a resume or a LinkedIn URL. I rank the live openings against it, rewrite the resume and write a cover letter for the best of them, and fill in the employer's form when you press the button. You read, you decide what goes out.
01Drop your resume
A PDF or a LinkedIn URL. About a minute, once.
02I rank the openings
Every weekday morning, the live catalog scored against your history. Up to 20 worth your time, not two hundred links.
03Each one is written up
Resume rewritten for the posting, a cover letter, a fit score. Press send, or let me fill in the form.
- New matches ranked and written before you are up.
- Every bullet stays inside what your history supports. Nothing invented.
- Queued, submitted, interviewing, offer: one screen, not a spreadsheet.
Free, no card. Up to 25 applications sent a week, and nothing is sent until you say so.
Listed from the job board Etched publishes. Refolk is not the employer and does not handle their hiring. Applications go to Etched directly.
Hiring at Etched? Post this role to the Refolk talent pool for $299, once. No commission, refunded if I find nobody. Post a role for $299