Machine learning engineer Jobs in Leverkusen

We are building the world's decision infrastructure: the strategic twin of every industrial organization for complex combinatorial problems. We unlocked what wasn't possible before: mapping reality with its levers and constraints into a computer, then running n-dimensional optimization on critical value-chain decisions. We help Europe stay strong and the German Mittelstand make good decisions between market shifts, orders, machines and people.

Founded by Dana (CEO) and Erik (CTO) from RWTH research, backed by a €2.8M round led by Vorwerk Ventures, with customers live on our optimization models.

We move fast. We care. No patience for problems left unsolved.

Your mission

Hi, I'm Erik, CTO of Circonomit. This ad is specific on purpose: you should be able to tell from it whether this is your job.

We build decision infrastructure for industrial companies: our customers model their production, with its capacities, costs and constraints, and we compute the answer to "what should we do?" before the decision is made.

Your mission: scale our data platform for multiple customers and domains. Customer data arrives in many formats from many sources, ERP exports included; it has to land in our system consistently, stay manageable at millions of records, and be fast to query when models and results are on-screen. The product is a data-heavy web app with a math-heavy backend (Python, TypeScript, Solid.js, Postgres, on Cloudflare): your center of gravity is data and infrastructure, but you touch the whole stack. We are looking for a system thinker who wants to shape the architecture.

We will not sugarcoat it: short deadlines, real workload, and a platform that grows while customers already run on it.

What you'll own

• The data platform end to end: ingestion, storage, and the query paths that keep millions of records fast on screen. You design it, you build it, you run it in production.
• The platform layer, directly with the CTO: infrastructure, deployments, and the path to production for the whole team.
• Performance and correctness work in the optimization engine, together with our research engineers.
• Your features from first line to production: nobody hands you a ticket and waits.

How we work

Small team, short lines of communication, no layers. You own your features end to end: you build them, you ship them to production yourself, you run them.

Feedback runs both ways, and it runs continuously, in daily work and in weekly one-on-ones. We talk as equals. We communicate proactively and flag it early when something isn't working out. Saying no is part of the job.

We like this work, and you'll notice. We review each other's code, and we like being together in the Cologne office, because the fastest conversations still happen in a room.

Requirements

• 5+ years building production backend or data-intensive systems.
• Strong Python and solid TypeScript, with the appetite to work in both daily.
• Analytics-data experience: real SQL depth, plus any of Parquet/Arrow/DuckDB/dbt/warehouse internals.
• Production PostgreSQL (multi-tenant, migrations, performance).
• Data pipelines over heterogeneous sources (ERP exports, Excel, CSV) onto object storage (S3 or comparable), built to keep working as customers and formats grow.
• Distributed systems and API design: typical failure modes, HTTP/auth/caching, sane versioning and error behavior.
• Engineering craft: design patterns, CI/CD pipelines you set up yourself, prototypes you can throw away.
• You have worked in a team, not mostly alone, and you know from experience which agile practices carry their weight.
• You have experience working with Linear, Jira, or a similar project management tool.
• You communicate proactively, too much rather than too little, and you organize yourself and your work.
• German at C1 or better, and fluent English. Team life runs in German; code and docs are English.
• NRW-based (Cologne office), optionally Munich, Stuttgart, Berlin area or else willing to work hybrid. Open to find a way if we fit.

Nice to have: a substantial migration you carried out end to end (system, stack, or data) · telemetry and observability · Cloudflare or comparable serverless infrastructure · frontend experience, ideally with Solid.js or another reactive framework · optimization/OR exposure · manufacturing or ERP data experience.

You are structured and biased for action, and you have shown you play to win wherever life has put you so far.

This role is not for you if you want a clearly defined box, or work handed to you in tickets. We are a small team, pre-Series A: structures and processes are still being built, by us and by you. Priorities shift with customers, and some weeks a deadline bends the plan.

Benefits

• Impact. You see your results in the real world, fast. The problems are concrete (production plans, capacities, logistics), and customers can measure in euros what your work changed.
• Ownership. What you build, you ship, and you own it once it runs. No layers between you and production, no approval chain that turns your decision into someone else's.
• The people next to you. You work directly with the CTO on the platform layer, and with the math and OR team on the optimization engine.
• Shaping. The processes are still forming, and whoever shows up with a better way of doing things gets to install it. This is a place for system thinkers: the interesting part is seeing how the pieces fit and where they will break. The company is early, and your scope will grow with it.
• Feedback speed. You will know where you stand. We say things out loud, we adjust, and we expect the same from you.
• High stakes. Competitive salary and relevant room in the equity package (VSOP) to match your contribution and your career development.
• The basics. Hardware of your choice · AI tooling budget · sports membership · Deutschland-Ticket.

Process: first conversation, 20 minutes, with me or with the engineer who owns our frontend. Second conversation with the two of us together; this is where the technical questions get deep. Then a hands-on challenge, about three hours of your time, followed by a 45-minute walkthrough where you explain your solution and the decisions behind it. Finally, the team: online first, then in person at the office. Both sides decide. Two to three weeks end to end, and you hear back within days after each step. Before you decide, ask us to put you on a call with someone who will tell you the good, the bad and the ugly.