Agentic Information Engineer (f/m/d)

Startdate

From 01.09.2026 onwards.

What you will be doing here

At MAGNOTHERM, we are developing a revolutionary refrigeration technology that has the potential to eliminate up to 30% of energy demand and 100% of direct greenhouse gas emissions in the cooling industry. Our vision is to provide the cleanest and most sophisticated cooling solutions to reduce climate impact and costs. Founded in 2019, our deep tech company is a spin-off of the Technical University of Darmstadt.

MAGNOTHERM is scaling up its engineering organization to build superconducting magnetocaloric cooling. As we grow, the company's information backbone must grow first — and we're building it, not buying it. You are that backbone: one person, amplified by AI agents, who does what traditionally took an entire IT department. You'll run our infrastructure, build our internal tools, make our company knowledge agent-accessible, and teach the team to work agentically. This role did not exist five years ago anywhere.

Curious to learn more? Then apply now!

What you'll own:

• Agentic toolchain (core mandate). Design and build the company's internal tool suite — PDM/BOM and engineering-change management, knowledge base, procurement & inventory, controlling & budgeting, test- and quality-data traceability with per-unit records, hiring pipeline, CRM, project & program tracking, asset & lab management with calibration schedules, and time & resource planning — as open-source + custom builds, developed with and operated by AI agents. Build-not-buy is the explicit company decision; you make it viable.

• Company knowledge as a system. Structure our documentation, engineering data, and decisions so LLMs can use them: RAG pipelines, MCP servers over internal systems, and custom agents/assistants the whole team queries daily. You own knowledge continuity — what we learn must outlive any single laptop or head.

• Infrastructure & IT. Classic sysadmin scope, run agentically: servers, networking, identity & access management, endpoint management, backups and disaster recovery, security — including an on-prem server as well as external resources for LLM inference. You are the human accountable for access control and production changes; agents do the routine work.

• Engineering-team leverage. Streamline the toolchains our engineers build and use — test-bench data pipelines, automated test reporting, live rig-telemetry dashboards, and simulation batch workflows. Harden the quick tools engineers write into maintained ones.

• Enablement. Every engineer gets an LLM budget; you make it pay off. Run internal training, build playbooks and shared agent configurations, and raise the whole team's agentic-engineering baseline.

• UX of internal tools. Internal tools people hate don't get used. You apply our design system and basic UX craft so the tools you ship are ones a production technician actually opens.

What you would ideally bring to the table

Must-have skills & experience

• Software generalist: strong Python and/or TypeScript, full-stack comfort (simple web apps, APIs, databases), 4+ years building real systems — internal tools, platform, DevOps, or product.

• Demonstrated agentic engineering: you've built agents that do real work — LLM APIs, tool use / function calling, at least one agent framework (Claude Agent SDK, LangGraph, or similar), and you can explain your eval and guardrail approach. A portfolio outweighs any certificate.

• Context/knowledge engineering: RAG over messy real-world documents, vector/hybrid retrieval, and ideally MCP server experience (scarce in the market — show us one you built).

• Linux system administration: servers, containers (Docker), networking, backup strategy, user/access management. Comfort being the last line of defense when something is down.

• Self-hosting and open-source fluency: you've deployed and lived with tools like ERPNext/Odoo, Outline/Wiki.js, Nextcloud, Grafana, or similar — and you know when self-hosting is a trap.

• Security fundamentals: least-privilege access, secrets management, and specifically agent security — scoping what an autonomous agent may touch, treating agents like employees with credentials.

• Working style: autonomous, product-minded toward internal customers, and able to explain technical decisions to non-software engineers.

• English language skills (C1+) and German language skills (C1+).

Nice to have

• Local LLM deployment (vLLM/Ollama, quantization, GPU server operation).

• UI/UX design competence beyond "makes it work" — Figma, design systems, information architecture.

• Experience as an early or solo IT/platform hire at a startup, or a homelab that got out of hand.

• Exposure to hardware-company data and workflows: CAD/PDM, BOMs and engineering changes, procurement and inventory, test data and per-unit traceability, ERP concepts.

• Teaching/enablement experience — workshops, internal docs, developer advocacy.

What makes this role unusual

• Founding scope: responsibilities that are normally split across IT leadership, platform engineering, and AI enablement sit in one role — with the authority that comes from owning all three.

• You define the discipline: there is no playbook for an agentic one-person information backbone; you write it, and it becomes how the company operates.

• Leverage as the KPI: your success is measured in how much faster everyone else moves, not in tickets closed.

As a startup, we value ownership, pragmatism, and a willingness to go where there's no playbook. This role is for people who love turning a blank page into a system the whole company relies on.

What you will get from the job

• Flexible hours and 30 days of vacation.

• Compensation aligned with startup standards, with room to grow as the company scales.

• A team that's sharp, interdisciplinary, and doesn't take itself too seriously.

• The chance to pioneer technology that actually helps the planet.

• Ownership, responsibility, and a culture where learning from mistakes is expected.

• A front-row seat as a deep-tech company scales.

• Memes. Good ones. And lots of them.

About us

We believe in science; we believe in facts. Climate change is speeding up more than ever and we need to do something about it. We are convinced that we need better technologies to overcome this tremendous challenge – or more precisely: this tremendous threat to humanity. We are on a mission to decarbonize the refrigeration industry and reduce its contribution of 7% to global greenhouse gas emissions.

Support us in making our vision a reality – join us!

We are material scientists, chemists, physicists, a lot of engineers as well as product and graphic designers. We are an agile team of bright minds with diverse and international backgrounds. We love hanging out together, we argue about new technologies, politics, the best sport, and food trends. Basically, we are human.

MAGNOTHERM is an equal opportunity employer. We embrace diversity and are committed to creating an inclusive environment for all employees. We are open to all groups of people without regard to age, color, national origin, religion, gender, sex, sexual orientation, gender identity and/or expression, marital status, or any other legally protected characteristics.

Do I need to meet all the requirements to apply?

Studies by several different sources have shown that, on average, men will apply for a job if they meet 60% of the application requirements. In contrast, women/non-binary people will seek to match a much higher percentage of the requirements before applying.

We encourage everyone to apply and give us a chance to evaluate your skills and experience. We are all learning on the job, and although the listing above has been carefully compiled, we are also open-minded and interested to hear about the value you can bring to the role and MAGNOTHERM.

How can I demonstrate that I have particular needs in the application process?

For people living with disabilities, chronic illnesses, or neurodiversity, adjustments and support can make a decisive difference in the application process. If you need any specific accommodations (tools, time, etc.) and feel comfortable disclosing this, please let us know.