AI-First Full Stack Developer

Most AI projects end in a slide deck.
Ours go live.

Lugano, Switzerland or remote within the EU

PHILIPP PLEIN is a Swiss luxury fashion group with several brands, stores and online shops around the world. We are AI-first: we rethink how the work is done, then build it with AI. What you build lands directly with our merchandise, allocation, finance and operations teams.

• Learn how our tech stack works end to end and find the weak spots: slow, costly, fragile or outdated.
• Fix those spots (speed, cost, stability, security, clean code), then keep them healthy.
• Build AI-driven solutions end to end, frontend, backend and database, from first idea to daily use.
• Build agents and web apps for business teams; connect product, PLM and sales data (read only).
• Run ETL pipelines with Airflow: gather data from all systems, clean it, check it, improve it.
• Track every app, agent and pipeline: who did what, errors, cost and quality, so you can fix fast and keep improving.

• Experience:
  
  3 to 5 years as a developer shipping software to production, at least one LLM solution used by real people, fluent English.
• Frontend:
  Type Script, React or Next.js; clear, fast screens for business users.
• Backend:
  Python (FastAPI) or Node.js; REST APIs, sign-in, background jobs, automated tests.
• Databases and data:
  PostgreSQL, strong SQL, data modelling, migrations, query tuning; ETL orchestration with Airflow, data quality checks, dbt, Snowflake or Databricks.
• LLMs and agents:
  Claude and OpenAI APIs, prompts with cost limits, structured output, MCP, multi-agent with human approval (Lang Graph, Pydantic AI or Claude Agent SDK), RAG, embeddings, vector search (pgvector), tool calling.
• Data science: pandas, scikit-learn, PyTorch, LightGBM or stats models; forecasts tested on past seasons.
• Governance and tracing: access rights, GDPR and Swiss data law, versioned prompts and models; audit trail (who, when, old to new); run tracing (Langfuse or similar), error tracking (Sentry or similar), cost, speed and quality per call, test sets for LLM answers.
• AI coding and delivery:
  Claude Code, Cursor or Copilot every day; you review, test and fix what the tool writes. Docker, Azure or another cloud, CI/CD with Git Hub Actions.
• Working skills: process mapping; turning what users say into clear specs and tickets.

• You start from the work, not the model.
• You ship things people use, then make them better.
• You check every number and every AI answer before you trust it.
• You explain your work in simple words, so anyone in the company can follow.

#AI Engineering #Full Stack #LLM #Hiring #Lugano #Remote Jobs