Senior Full Stack Engineer (AI, Document Processing) - German-speaking

Now we're adding a senior engineer to the core product team. This role is for an engineer who can ship end-to-end, own production outcomes, and move fast without lowering the bar. You'll be part of a small, talented, highly driven dev-team all of whom bring the same level of ownership, curiosity, and ambition you value yourself. No passengers. No layers. High trust, high ownership, high standards.

What you'll do

• Own end-to-end features across our Next.js frontend and FastAPI backend-design, build, test, ship, and monitor.

• Harden and evolve our document processing pipeline: uploads * S3 * async processing (Azure Document Intelligence / Textract / multimodal) * DB persistence * semantic search (vector stores).

• Improve reliability and correctness of event-driven processing (SQS * Lambda handlers), idempotency, retries, DLQs, and observability.

• Build and maintain real-time UX (WebSocket events + SSE streaming) for long-running AI and document workflows.

• Work in the billing system (Stripe subscriptions, consolidated subscriptions, plan switches, schedules, proration) and make changes safely via tests + rollout strategy.

• Collaborate directly with users/stakeholders (often in German) to understand workflows and iterate quickly.

• Raise engineering standards through crisp PRs, tests, refactors, and documentation.

Tech stack (what you'll touch)

• Frontend: Next.js 15, TypeScript, Redux Toolkit + RTK Query, MUI + Tailwind, NextAuth (Cognito), WebSockets, SSE.

• Backend: FastAPI (async), MongoDB (Motor), Pydantic models, repository/service patterns.

• AWS: ECS (containers), Lambda, S3, SQS (+ DLQ), API Gateway / NLB, Secrets Manager.

• AI: multi-provider AI module, OpenAI Agents SDK (tool use, streaming, context management), extraction into typed Pydantic models, evaluation/quality workflows., * Stripe subscription lifecycle experience (schedules, proration, consolidated subscriptions, webhooks).

• OCR/document extraction pipelines and working with imperfect inputs (PDFs, scans, tables/forms).

• Building agentic systems: tool design, guardrails, evals, prompt versioning, cost/perf tradeoffs.

• Comfort owning a service: operational runbooks, alerts, and "never fail the same way twice" postmortems.

• Experience with MongoDB Atlas Search and vector search in production.

How we work (and what we value)

• High ownership: you ship outcomes, not tickets.

• Type safety and correctness: Pydantic models, enums over strings, explicit error handling, tests on critical paths.

• Move fast with discipline: small PRs, measurable changes, and production awareness.

• Thoughtful use of AI tools: we use AI daily to accelerate; you must be able to verify, test, and own the final result.

First 30 / 60 / 90 days (success looks like)

• 30 days: run the stack locally; ship a few small production changes; understand our critical flows (billing, document processing, SQS/Lambda tasks, real-time streaming).

• 60 days: own a medium-sized end-to-end feature; improve reliability or testing in a critical area (billing or document processing); be comfortable taking primary responsibility for a production incident in your area.

• 90 days: lead a high-impact project (e.g., quality improvements for a real PDF workflow, billing lifecycle hardening, or real-time reliability) and set a sustainable standard others can follow.

Logistics

• Remote (time zone overlap with Central Europe preferred).

• German-speaking required.

• Compensation, engagement model, and ESOP depend on seniority and scope; we hire for impact and long-term ownership.

Do you have experience in TypeScript?, * Have 5+ years building and shipping production web systems and can independently deliver end-to-end features.

• Are strong in TypeScript/React/Next.js and comfortable with complex client state (RTK Query caching/invalidation, auth/session flows, real-time updates).

• Are strong in Python/FastAPI with async I/O, API design, and disciplined data modeling (Pydantic, enums, typed interfaces).

• Have shipped systems on AWS and understand queues, retries, idempotency, and failure modes in distributed systems.

• Can debug production issues quickly (logs, metrics, traces), write fixes safely, and communicate clearly.

• Have a product mindset: you enjoy talking to users and see "product" as part of the job.

• Speak German fluently (you can run user calls, write clear German messages, and work with German legal/administrative terminology). English is used in code and many docs.

NotarNow is building a digital notary platform that streamlines German document and notarization workflows. We combine a modern web product (Next.js/TypeScript) with a FastAPI backend deployed on AWS and ship AI-powered document understanding, search, and real-time collaboration. We're one of Europe's fastest-growing bootstrapped legal-tech startups. We hit €1.5M ARR in under two years and are scaling fast. Our products run real high-stakes professional workflows, process real documents, and are used by 1,000+ legal professionals.