AI Engineer

Join Crypto Finance's Data & Analytics team as an AI Engineer to shape the company's AI strategy and execution. You'll identify high-impact AI use cases across the business, engineer production-grade AI and automation solutions, build the firm's AI governance practice, and contribute to the data platform that underpins Finance, Operations, and Risk reporting. This hybrid role in Zurich requires someone equally at home in a stakeholder workshop and an IDE, pushing back on bad use cases as much as shipping good ones.

What you'll do

• Work with stakeholders across Compliance, Trading, Operations, Legal, Sales, and Finance to identify, scope, and prioritize use cases that move the needle.
• Engineer production data solutions including deterministic automations, AI agents, RAG systems over internal documents, and structured extraction pipelines.
• Build the firm's innovation lab environment where new use cases can be prototyped and evaluated.
• Maintain prompt and skill libraries as reusable, version-controlled assets.
• Maintain the firm-wide inventory of AI systems and use cases, including those built outside Data & Analytics.
• Run the operational side of the company's AI approval process: documentation, risk classification, model cards, and evaluation artifacts.
• Conduct technical review of new AI initiatives proposed elsewhere in the company and advise on scope, risk, and design choices.
• Contribute to ELT pipelines on the Dagster, SQLMesh, and dlt stack, particularly where AI and automation use cases require new data sources or transformations.
• Build the data substrate that AI workloads consume, including feature views, document indexes, and structured event tables.
• Maintain infrastructure as code in Git with proper review and deployment standards.

What you bring

• 3-6 years of relevant experience: shipped AI engineer, analytics engineer with AI focus, or data engineer pivoted to AI work.
• Demonstrable production experience with LLM applications, including at minimum structured extraction with LLMs and agentic patterns such as tool use and multi-step workflows.
• Strong Python; comfortable building production-ready code, not just notebooks.
• Working knowledge of evaluation discipline for LLM applications (eval sets, regression tests, observability), conceptual understanding of retrieval, and how to handle hallucination.
• Familiarity with at least one orchestrator (Dagster, Airflow, or Prefect) and one transformation framework (SQLMesh or dbt).
• Solid SQL including window functions, joins, and query design.
• Cloud experience, ideally GCP and BigQuery.
• Genuine curiosity about regulated environments and the discipline they require.
• Professional proficiency in English.
• Eligibility to work in Switzerland: a Swiss work permit or EU/EFTA citizenship.
• You reach for the simplest thing that works. Boring SQL before vector search. Rules before agents. You can explain why.
• You are comfortable saying "this is not an AI problem" when the right answer is a dashboard, a process fix, or a deterministic script.
• You can run a stakeholder workshop and write production code in the same week.
• You take documentation, evaluation, and audit trails seriously, not as overhead.

Nice to have

• Prior exposure to regulated financial services, crypto, or comparable model-risk environments.
• German language proficiency.

What we offer

• A modern and central workplace in Zurich with top-tier infrastructure.
• Opportunity to shape the future of finance at the cutting edge of blockchain and fintech.
• A collaborative, entrepreneurial culture with flat hierarchies.
• Meaningful responsibility with room for learning and professional growth.
• Regular company-wide meetings, knowledge-sharing sessions, and team events.

About Crypto Finance

Crypto Finance is an international blockchain and fintech company headquartered in Zurich's Prime Tower, offering B2B digital asset solutions. The team is led by experienced professionals and values innovation, continuous learning, collaboration, and an entrepreneurial spirit with flat hierarchies.