Senior Ruby on Rails Engineer

About Amba

Amba builds remote monitoring for care settings. Our sensors and alerts tell a care team something is wrong before it turns into an emergency. The product is a Rails 8 monolith, built on PostgreSQL and Sidekiq, using ViewComponents and RSpec, and covering multi-tenant billing and medical device integrations.

We're hiring two senior Rails engineers in the US to start our US team.

How the Work Gets Made

• AI agents write most of our code. Every piece of work starts as a plan on a GitHub issue, and a human approves that plan before an agent builds anything.
• Agents work test-first and stop at a draft pull request. Usually two humans review it, and anything touching migrations or tenancy boundaries gets a stricter review again.
• Some work is attended, with a developer alongside the agent and review points as they go. Unattended work, where the agent builds the approved plan alone, gets the closest review of all, because nobody was watching it happen.
• No agent has ever merged code here, and none ever will. Our alerts affect someone's care, and these gates are what let us keep pace without taking risks with it.

Your Week

• Meetings are kept out of the way: the team aligns on Monday and posts a written update on Friday. Standup is async the rest of the week.
• We work in eight-week cycles, split into four two-week sprints with natural breaks between them.
• One engineer covers support requests for a whole cycle so nobody else gets pulled off their work. Six of the eight of us share that rotation, so a turn comes round about every six cycles, and less often as the team grows.

Your Day

Most days are a mix of:

• Scoping features and writing the development plans that agents build from.
• Steering agents when they need it.
• Reviewing pull requests from the rest of the team.
• Bringing your own ideas. Product ideas start in an open Slack channel anyone can post to, you shape the code you work in, and each cycle ends with a retrospective where the team changes how we work.

You will write code too, just less of it than you do now.

What We Need

• Rails depth you use for judgement: you can read a plan and spot the migration that will lock a table, or the query that skips a tenancy check.
• Real Claude Code experience: instruction files in repos you have shipped, custom commands, subagents, MCP. Cursor and Codex count for something, but Claude Code is what we run on.
• Enthusiasm for working this way. We'd rather hire someone who finds this interesting than someone who will put up with it.
• The ability to work on your own. You would be among the first people we have hired in the US, with about four hours of daily overlap with the rest of the team. Most of your day, you set your own direction.
• Enjoying the human side of the role: for part of each day you will be the only engineer US colleagues can reach, so questions from sales and customer success will find you. We want people who like explaining things to non-technical colleagues and digging into a customer problem, not people who tolerate it.
• Able to work from Eastern or Central US time.
• Existing US work authorisation. We cannot sponsor visas

Not a Fit If

• You love writing every line yourself. Plenty of excellent engineers do, and they would be unhappy here.
• You'd relax when reviewing an agent's work because the tests were green. We do the opposite: the less a human watches the work happen, the harder we look at it.
• You want to be left alone with the code. Part of this job is being the friendly face of engineering for US colleagues.
• You want someone checking in on you throughout the day.

The Team

Eight engineers, fully remote, none of them currently in the US. These two roles start our US team. You would report to Paul, our team lead. We need you in Eastern or Central time, since that's what makes the four hours of overlap with the UK real rather than theoretical. Pacific time gives us two, and we have watched that grind a team down before.

We do not run an on-call rotation today; the support cycle above is how we cover things. We may add one as the team grows, and we'd design it with the team rather than announce it.

Compensation & Benefits

Component

Detail

• Base Salary: $150,000 – $165,000 based on experience
• Medical Insurance: Coverage begins after 30 days, with company contribution
• 401(k): Immediate eligibility · 100% employer match on first 4%
• PTO: 15 days, rising with tenure + company-approved holidays
• EE Bonus Plan: Eligible for company wide bonus scheme
• Equipment: Laptop and home office setup provided
• Travel: Rare, and only with plenty of advance notice

How We Hire

We only consider applications that show evidence of working inside a strict AI workflow. Send whatever makes that case best: a repo, an instruction file, a writeup, a video, anything you think proves it.

• A short written exercise. We give you a problem and you rough out how you would scope and phase it. No code, and it should take about 30 minutes.
• A conversation with Paul, the team lead you would report to.
• A 30 to 45 minute conversation with our CTO about your experience and your exercise.

Then an offer. We do not set leetcode puzzles.

How to Apply

With your application, tell us about a plan you wrote for an agent that went wrong, and what you do differently now. Every one of us has an answer to that question. If you do too, we should talk.

Via LinkedIn or directly via workwithus@amba.co