MISSION

The intern will join our Process Optimization Team for a six-month assignment contributing to live AI projects and exploratory research. They will work directly alongside senior advisors and project leads on production AI agents already deployed across the business, as well as on proof-of-concept work that shapes our next wave of AI initiatives.

MAIN ACTIVITIES AND RESPONSIBILITIES

• Extend and harden internal AI agents already in production, including a procurement email agent operating in O365 with read-only actions on PO status, invoice, and payment queries.

• Develop new AI use cases under exploration, including trade confirmation matching, remittance application, and other workflow automation candidates.

• Build and integrate tools using LLM platforms (Claude, OpenAI) via API, MCP connectors, and OpenAPI Actions; design prompts, tool schemas, and authorization layers.

• Implement security and governance patterns: prompt injection defense, idempotency, audit logging, sender allowlisting, and circuit breakers.

• Run structured evaluations, including adversarial test cases, and document results.

• Contribute Python tooling for data pipelines, file format conversions, and integration with enterprise systems (Microsoft 365, Azure, internal APIs).

• Produce clear technical documentation and present proof-of-concept outcomes to IT and business stakeholders.

REQUIRED EDUCATION AND EXPERIENCE

• Master's student or recent graduate in Computer Science, Data Science, AI, or a related quantitative field.

• Hands-on experience with at least one LLM platform (Anthropic Claude, OpenAI, or equivalent), including prompt design, function/tool calling, and structured outputs.

• Eligible to undertake an internship in Switzerland.

QUALIFICATIONS AND SKILLS

Required

• Strong Python skills (data structures, async, REST APIs, error handling).

• Working knowledge of cloud and identity concepts (Azure, Entra ID, OAuth) sufficient to reason about enterprise integration.

• Comfortable working in English and French in a professional setting.

Preferred

• Exposure to MCP (Model Context Protocol), OpenAPI Actions, or agent frameworks.

• Experience with Microsoft 365 / Graph API, or with enterprise email and document workflows.

• Familiarity with security topics relevant to AI agents (prompt injection, tool-layer authorization, audit logging).

• Prior internship, project, or open-source contribution demonstrating shipped code.

• Full professional proficiency in spoken and written English and French is required; knowledge of other languages is an asset.

• Swiss national, EU national, or holder of a valid Swiss work permit.