Senior Backend & Infrastructure Engineer

Wasteer GmbH is hiring one engineer to own the backend, the databases, and the cloud and edge infrastructure that every other service depends on. Berlin, three days in the office and two from home.

What we do

Wasteer GmbH is hiring one engineer to own the backend, the databases, and the cloud and edge infrastructure that every other service depends on. Berlin, three days in the office and two from home.

What we do

Wasteer puts computer vision and machine learning on the waste stream at European incineration plants. Models watch the tipping floor and the crane, classify what is actually in a load, flag contamination the operator would otherwise miss, and turn all of it into the record that gets billed, audited and reported to regulators. Inference runs at the edge, on live industrial hardware, in real time, and has to be right often enough that people trust it over their own eyes.

Wasteer is post-revenue, with paying customers running plants that handle 16 percent of Europe's waste. What you build here is in production on day one. When we go down, a plant goes back to paper.

What you own

You own the systems everything else depends on: the FastAPI backend, the PostgreSQL databases behind it, our GCP footprint, and the Cloudflare edge. Right now one person carries all of that. You would be the second, and over time the person who carries it. This is not a ticket queue. If the API is slow, a migration is risky, or the cloud bill doubled overnight, it is yours to figure out and fix.

What you will actually work on

• The main backend: Python and FastAPI over PostgreSQL, with authentication and role-based access control across the API.
• The database: schema design, migrations that run against live facilities, query performance, data retention.
• Google Cloud: containerised services and scheduled jobs, managed Postgres, object storage, secrets. Docker, not a managed platform that hides things from you.
• Cloudflare Workers at the edge, with the infrastructure managed as code.
• Keeping the services around all of that reachable, observable and cheap. Some of them talk to hardware at the plants. You will not be writing the computer vision, but you will be the reason it can authenticate, store its results and stay up.

What we need

Must have

• Strong Python. You have shipped and maintained a production API, not just scripts.
• Real SQL and PostgreSQL. You can read a query plan, add the right index, and write a migration you are willing to run on a Friday.
• Production cloud experience, ideally GCP. You have been the person who got paged, found the cause, and fixed the cause.
• You can operate without a spec. We tell you the outcome. The approach is yours.

Nice to have

• Cloudflare Workers, Terraform, or edge and IoT systems.
• Having worked somewhere small enough that \"who owns this\" had an obvious answer: you.
• You founded something technical, whether or not it worked out. Ex-technical founders are explicitly encouraged to apply: this is the same ownership, without the fundraising.

Not a fit if

• You need a defined scope, a backlog groomed for you, and a clear boundary between your job and someone else's.
• You want to spend the first three months rewriting what exists before shipping anything.
• You need every decision reviewed before you make it.

What it is really like

We are small, so the whole surface is yours. The decisions you make in your first months are the ones this company runs on for years, and you will see your work live at a plant the same week you write it.

The code is not uniformly clean. Some of it was written fast because a facility needed it working that week, and that was the right call. We refactor when it pays for itself, not because a pattern offends us. If you cannot ship a working, slightly ugly thing on Tuesday and improve it a month later, you will be unhappy here.

How we work

Simplicity of execution is the art here. Anyone can build the complicated version. We rate the solution a new person understands in a minute and nobody has to babysit. The best work here is the service nobody has thought about in six months, because it simply works.

Scalability is key. What works at one facility has to work at forty without a rewrite. Every design decision gets asked that question.

For an unproven idea, quick and dirty wins. Ship the rough version this week, put it in front of a plant, and let reality tell you whether it deserves the clean one. Spending months perfecting something nobody ends up using is the most expensive mistake we can make. Once the idea is proven, then it gets built to last.

Some things are not negotiable. Anything running at a plant has tests, and they run against a real server rather than a mock. Errors are handled explicitly and never swallowed silently. No hardcoded secrets, ever. Services stay small and clearly owned. You deploy your own work, because there is no separate ops team to hand it to.

How we hire

• Screening. We read every application ourselves. No recruiter filter and no keyword match.
• A project. We give you a real problem, close to what we actually work on, and you build it. No trick questions and no whiteboard.
• A walkthrough. You show us what you built and we ask why: the call you made, the one you rejected, and what you deliberately left out.
• Offer.

On the project, we look at how far you thought in both directions. Up: what this does to the system, the cost, the operator at three in the morning, the migration, what happens when it fails. Down: the detail you got right because you went and checked instead of assuming. A thorough, complete approach beats a minimal one that technically passes. Take the extra mile and tell us where you stopped and why.

Public git history, open source you maintain, and having built your own startup all count in your favour. Send the links.

First 90 days

• Day 30: you have shipped backend changes to production on your own, and you can explain how a load arriving at a plant becomes a classified, billable record.
• Day 60: you own at least one service end to end, including its deployment and its failure modes, and you have made one infrastructure decision we did not have to review.
• Day 90: someone other than you can take something you built when it breaks, because you documented it and it does not surprise people.

The practical details

• Title: Senior Backend & Infrastructure Engineer
• Location: Berlin. Three days in the office, two from home. Subject to change.
• Language: The team works in English. Any other EU language is a plus.
• Reports to: Fatih Unver, CTO
• Compensation: Discussed in the first call. We will not post a range we cannot honor.
• Start: As soon as you can.