copperhead is an open source agent that designs, documents and verifies real circuit boards. This role builds the infrastructure underneath the agents: the execution loops, sandboxes, job pipelines, evaluation harnesses and tracing that decide whether an agent's work can be trusted. The questions below are about systems you have built and operated, what you personally owned and what happened when they failed. Those are the ones we read first.

About copperhead

copperhead is an open source agent that designs, documents and verifies real circuit boards.

About the role

You’ll build the infrastructure underneath the agents.
The work spans execution loops, sandboxes, job pipelines, evaluation harnesses and tracing.
The goal is to make the agent’s work trustworthy in production.

Application notes

Please answer the role-specific questions field for field.
We read your examples of systems, ownership and failure handling first.

Name *

Email *

Phone (Optional, we may call you)

Can you work in person from Bengaluru? *

I am in Bengaluru

I am elsewhere in India and will relocate

I am outside India and will relocate

I cannot work in person

GitHub *

Portfolio, site or writing

Resume *

PDF or DOC/DOCX, max 10MB.

When could you start, and what notice do you owe? *

What are you expecting to be paid? *

Anything else we should know?

Optional, 60 seconds: what is a hard engineering problem you solved recently?

Which of these have you built or operated in production? *

Background workers or job queues

Message queues or distributed job processing

Containerised or sandboxed execution

LLM tool calling or structured outputs

Tracing, logging or observability pipelines

None of these

The hardest infrastructure, AI agent or developer-tooling system you have built. What did it do, and what did you personally own? *

A difficult technical failure in a system you ran, and how you solved it. What broke, how did you find it and what changed afterwards? *

Describe an agent or execution loop you built that called tools and had to recover from failures. How did it decide it was done, and how did it recover? *

How have you isolated untrusted or compute-heavy work? Sandboxes, containers, workers: what did the boundary look like, and what got through it? *

How have you measured whether a probabilistic system was getting better or worse? Describe the evaluation, what it caught and what it missed. *

A repository, pull request or system you would point at as your best work *

Describe a technical problem you owned end to end when the requirements were not clear. What did you decide, and what did it cost? *

What is the most unfamiliar technical domain you have had to get productive in, and how long did that take? *

Any exposure to electronics, hardware or EDA tools?

Made withBinary

Founding Software Engineer - AI Infrastructure, copperhead

About copperhead

About the role

Application notes

Founding Software Engineer - AI Infrastructure, copperhead

About copperhead

About the role

Application notes