Senior Fullstack Engineer (m/f/d) | Python, React | LLM Security Platform | Düsseldorf

Stellenbeschreibung:

YOUR MISSION

As Senior Fullstack Engineer, you're owning the development of our AI platform that automates the architecture-level security reviews our consultants currently do by hand. You spend most of your time on the backend, building and extending the AI-integrated services, async pipelines and APIs that power our research platform. The rest is the interface layer, where consultants interact directly with the AI's output.

You’ll be part of a a small, tight-knit setup where every contribution counts. Whilst stakeholders communicate their requirements, you know how to put them into practice.

YOUR ROLE

Backend & API : Build and extend FastAPI endpoints, Celery task orchestration, PostgreSQL schemas, and GCS object storage; clean integration with the AI/ML pipeline.
Data Flow Ownership : Design how the platform's data is stored, versioned, and served to the UI.
Knowledge Approval Workflow : Design and extend the lifecycle layer for documents in the platform: data model, SLA tracking, escalation signals, and provenance metadata.
Consultant-facing UI : Build the surfaces where AI output meets users: long running task UX, real time updates, and the path from polling to WebSocket/SSE streaming.
End-to-End Delivery : Work with the AI team and an external UX designer to ship features from pipeline to UI – and own what works under the hood.
Platform Reliability : Shape test strategies, structured logging, performance budgets, and the SDLC and agile practices the team needs as it grows.

YOUR PROFILE

Must haves

Fullstack experience : You bring 5+ years with recent production ownership of complex, enterprise-level software, ideally including AI-integrated workflows you've taken from pipeline to real users.
Python and FastAPI at senior level : You design APIs, validate requests with Pydantic, handle errors cleanly, and integrate confidently with our AI/ML pipeline.
PostgreSQL : You know your way around schema design, migrations, query optimization, and audit/versioning patterns.
Async at scale : You're comfortable with streaming responses (SSE/WebSocket), bounded concurrency, retry strategies across nested boundaries, and idempotent task design with Celery or equivalent.
Vector search : You have working knowledge of Weaviate, pgvector, or similar.
LLM production experience : You've shipped features where LLM output flows end-to-end through real pipelines to real users in a live production environment, including streaming responses, structured output handling, retries, and graceful recovery from non-deterministic behaviour.
React and TypeScript : You build data-driven UI for domain experts using state management like TanStack Query or Zustand, and you understand that AI/LLM product UX is fundamentally different from CRUD UX.
Production deployment : You're at home with Docker, Kubernetes or equivalent, CI/CD, and observability tooling like Sentry or OpenTelemetry.
End-to-end ownership : You've driven complete features or systems from architecture through to production, bringing a rounded understanding of how a platform fits together as a whole.
Product judgement : You read a feature request and know whether it fits where the platform is going, and you speak up early, shaping engineering and design decisions to keep the consultant workflow strong.
SDLC and agile practices : You've helped build these in a small team.
English communication : You communicate confidently and clearly in English, spoken and written.
GCP experience : Cloud Run, GCS, Secret Manager, Vertex AI.
Graph databases : You've worked with Neo4j or similar alongside vector search, or you're curious to pick them up fast.
AI/LLM eval tooling : You're familiar with eval frameworks, prompt versioning, or regression testing on non-deterministic outputs.
Security engineering : You know authn/authz patterns, session management, secrets handling, or you're familiar with OWASP, threat modelling, and MITRE ATT&CK.
Test strategy : You've helped build testing practices systematically within a team.
Domain background : You've worked in developer tools, data-heavy SaaS, or expert-facing products.
German : You have language skills at A2+ level.

WHY CLOUDYRION

30 days of vacation plus additional days off on Christmas, New Year’s Eve, and your birthday
Hybrid work with two remote days per week, trust-based working hours and the opportunity of Workations in any EU country
Individual learning and development opportunities, including full coverage of advanced trainings, certification costs, and educational leave for exam days
In-house gym, massage chair, gaming area, and dedicated break spaces
Regular team events and plenty of space for real collaboration

ABOUT CLOUDYRION

CLOUDYRION was founded in 2020, bootstrapped and built around a simple idea: cybersecurity done properly, by people who actually care.

We work with large enterprises in the telecoms, banking, insurance, automotive and critical infrastructure sectors. Our Secure by Design, Ethical Hacking and Secure Strategy Consulting services help them get security right from the ground up.

With our founder as CEO and free from investor influence, we are still upholding our values five years on: Fail fast and learn quickly. Be a ray of sunshine. Tear down those walls. Securing more than symstems.

Working at CLOUDYRION means being part of a team who love what they do, embrace diversity and celebrate their successes together.

This approach has proven its worth: our team now consists of 35 people and generates an annual turnover of around €6 million.

Learn more

Insights from real-world projects, technical deep dives and our take on secure-by-design, cloud security and modern security architectures can be found on our blog.

Authentic insights into our office life can be found on Instagram.

Do you feel inspired to take your next step? Then we look forward to your application. Whether classic or creative is up to you. What matters most is that you show us what drives you.