(Senior) Platform Engineer (f/m/d)

Most AI platforms break when they scale. We're building one that doesn't. Your mission: make our architecture a competitive advantage in reliability, performance, and operational clarity.

Our Mission

Our mission with MeinGPT: be the AI use case platform for the Mittelstand — the system that gets companies from their first AI experiment to AI running in their core processes.

MeinGPT already serves 50,000+ users from 500+ companies. What comes next isn't just more users. It's more depth: AI that connects to real data, real systems, and real workflows. That demands an architecture that doesn't get in the way. Platform Engineers build the foundation that makes it all possible — reliably, securely, and at any scale.

Our platform engineering challenges for 2027:

• Performance, cost, and observability: We measure, improve, and measure again: LLM streaming latency (directly visible to users), AI API costs per request, database query patterns under load, sandbox performance for code execution. The goal is a system that tells you what's going wrong before users do, and where you can deploy with confidence at any time of day.
• AI infrastructure: Routing AI requests reliably across six LLM providers with fallback logic, rate-limit handling, and cost tracking. Integrating with our meeting bot infrastructure for recording meetings across Google Meet, Teams, and Zoom. Operating the code execution sandbox for task automation.
• Architectural foundations: The platform has been shaped by years of real production use: well-understood, battle-tested, and still actively evolving. Platform work here means exercising architectural judgment: knowing when to add structure, when KISS is right, and how to continuously raise the quality of the systems everything else builds on.

Your Mission

AI products integrate, move, and process more data per user action than most SaaS: external sources synced continuously, context assembled in real time from multiple systems, streaming responses held coherent across distributed services, LLM requests routed with fallback and cost logic across providers. Getting this right, under real load, for real customers, is an architecture problem. Platform Engineers build and own that foundation.

The work is proactive: you read the system's signals (latency distributions, cost curves, error rates) and act before they become user problems. The goal isn't to keep pace with growth. It's to build a platform that's more capable and more reliable at 100,000 users than it is today.

What We Are Looking For

• 3+ years TypeScript in production backend systems — your focus is clearly on the backend, but you're comfortable in the frontend too and can jump in when needed
• You think about systems under sustained load: how they degrade, what breaks first at 2x traffic, where the cost curve goes — and you design for that, not just for today's requirements.
• Strong system design knowledge: PostgreSQL, APIs, queues, caching. You know when to use what and why.
• Experience with containerized deployments (Docker, Kubernetes) and the operational practices that come with them.
• When you see a latency pattern, cost spike, or error rate trend, you track it to its root and fix it properly — not apply a patch and call it a day.
• When data is missing for well-founded decisions, you add the instrumentation yourself.
• You share what you know (context, tradeoffs, open questions) and make the engineers around you better at their work.

What's in it for You

• Shape the technical foundation of a platform that's actively scaling: your architectural decisions outlast any single feature.
• Work in a novel, data-intensive field with interesting scaling challenges: making terabytes of data available to AI on-demand, providing compute infrastructure for AI agents at scale, and making MeinGPT a flexible platform to build AI applications on.
• Flexible working hours with the option to work from home 2 days per week.
• €1,000 annual learning budget, EGYM Wellpass (10,000+ sports facilities), Deutschlandticket.

How We Work

• We use AI tools (Cursor, Claude Code etc.) daily to move faster, but you stay in control and responsible for your code.
• Meetings are kept minimal to protect deep work time.
• Healthy Ambition: big goals, sustainable pace, people who actually like working here.
• Fast Delivery, Close to the User: we ship fast, stay close to our customers, and let direct feedback drive next steps.

Our stack

• Core: TypeScript, Go, tRPC, PostgreSQL, Kubernetes, Docker
• AI & Dev Tools: Cursor, Claude Code, CodeRabbit, Linear, GitLab, Notion
• AI Services: OpenAI, Gemini, Claude, Mistral, Gladia

What Happens Next

• Application: Submit your profile through our short form.
• Intro Call (30 min): A short chat about your motivation and our mutual expectations.
• Technical Interview (Remote, 90 min): A live coding session to showcase your skills.
• Founders Interview (On-site, 90 min): Meet the founders at our office and discuss your experience and ambitions.
• Offer: If it's a great fit for both sides, we'll make you an offer.