Data & AI Engineer (all genders), PharmaHera Service GmbH

Deine neue Umgebung: Behind PharmaHera is Atida - one of Europe's leading online pharmacy groups, serving over 5 million customers across 9 countries. Together with the CEPD Group, we combine more than 25 years of pharmaceutical expertise with digital excellence to redefine the online pharmacy experience: We are building a trusted, personalized health ecosystem that guides people from prevention to cure - backed by medical expertise and driven by our conviction: Pharmacy is about people. Based in Berlin, we serve as the group's competence center for Germany and the Netherlands. As a specialized partner, we manage e-commerce, marketing, IT, and logistics for our local mail-order pharmacies, such as APONEO. Our European markets run on a single, standardized data platform (our Foundation layer). Your mission for the future: To transform this data infrastructure so it is not only usable by people, but specifically prepared for AI agents that read, analyze, and autonomously act on it. The role Reporting to the Data & AI Engineer Lead, you will own the data pipelines, datasets, AI agents and tools built on top of Foundation, and you will build most of them with AI agents as part of your toolchain. The role sits one level of abstraction above the code: you specify, review and take responsibility for systems rather than typing every line. Core responsibilities Pipeline development: build and optimize data flows in Apache Airflow, from our European markets into BigQuery. Modeling in BigQuery: implement tables and views in the Foundation layer, guaranteeing KPI comparability across Spain, France, Italy and Germany. Semantic layer and data contracts: keep business definitions consistent across dashboards, models and agents, and define the schema, granularity and quality a dataset guarantees. AI-assisted delivery: use Claude Code and the Anthropic API as your everyday engineering environment -- turn a requirement into a precise specification, let the agent produce the pipeline, then review, test and own the result. Building for agents: expose Foundation through MCP tools and skills so that AIDA and our decision agents can query it reliably. A dataset is finished when an agent can use it without being misled. Evaluation: build the evals, checks and monitoring that tell us whether an agent's output is correct. Unit tests are not enough for non-deterministic systems. Data products: design data products for management, finance, commercial, operations and marketing -- reusable across markets. Visualization and reporting: work with the business team on reports and consumable datasets in Apache Superset. Code and traceability: clean Python and SQL, with every task traceable through our GitLab-Coda integration. Agent-generated code included. Your Profile & Tech Stack What You Bring to the Role (Must-Haves): Education & Experience: 2+ years of hands-on experience in Data Engineering or Data Science. AI-Assisted Engineering: Regular, demonstrable use of an LLM coding agent (Claude Code or equivalent) in daily work, plus the technical judgment to critically review agent-generated code. Tech Stack & Cloud: Solid experience with Google Cloud Platform (especially BigQuery), Apache Airflow, strong SQL, data-oriented Python, and version control via Git/GitLab. Languages: Fluent English for daily collaboration within our international team. Mindset & Work Ethic: Strong analytical skills, an independent and results-focused approach ("you get the problem and context, the solution design is yours to propose"), and a commitment to deeply understanding code rather than using AI as a blind shortcut. What Rounds Off Your Profile (Nice-to-Haves): Experience with MCP (Model Context Protocol), LLM APIs, ClearML, Docker, or Apache Druid. Know-how in evaluating non-deterministic systems and using BI tools (ideally Apache Superset). Industry experience in the healthcare, pharmacy, or e-commerce sectors, plus basic knowledge of German. What we are not looking for Someone who ships code they cannot explain. Someone who needs a ticket with the solution already inside it or with full requirements in place. We act fast: you get the problem, the context and access -- the design is yours to propose. Someone who uses AI as a shortcut and doesn't understand the problem they are solving. Someone who only thinks short-term. We build future-proof solutions.