AI Senior Software Engineer – LLMs, Agents & Integrations

About us

As Switzerland’s leading digital hub, we provide our media and platforms with ideal framework conditions, support them through investments in technology, and create space for them to develop individually. We stand for interdisciplinary collaboration, innovation, and dynamic development.

We are on the move – and want to keep moving. We are farsighted. We are proactive. We are courageous. We are TX. TX Services is a part of TX Group.

About the job

We develop internal and external AI applications that support employees and customers, make corporate knowledge accessible, and simplify and automate work processes. To achieve this, we combine Large Language Models with corporate data, existing applications, and intelligent automation.

We are looking for a versatile Software Engineer with a deep understanding of LLMs, agent systems (agentic harnesses), and their practical integration. You will work closely with project leads, business stakeholders, and external developers, accompanying various initiatives from technical concept to prototyping and production deployment.

What would you do?

• Design and develop AI applications, assistants, and automations.

• Integrate LLMs, knowledge sources, and enterprise systems via APIs, SDKs, and the Model Context Protocol (MCP).

• Develop agents with tool access, context management, persistent state, and governed approval workflows.

• Connect backend services, data processing, and web interfaces into end-to-end solutions.

• Effectively leverage prompting, structured outputs, tool calling, and Retrieval-Augmented Generation (RAG).

• Evaluate models and technical approaches based on quality, data privacy, cost, latency, and operational requirements.

• Secure applications through testing, LLM evaluations, monitoring, and traceable error handling.

• Shape architectural decisions, code reviews, and technical documentation together with the team.

What should you bring?

Our existing stack forms an essential foundation of the role. We expect hands-on experience in the following areas:

• Languages & Software Engineering. Strong proficiency in TypeScript and JavaScript, strict typing, asynchronous processing, and modular architecture. Practical experience with at least one additional language, such as Python, and the ability to work cross-technology.

• LLMs & AI Integration. Vercel AI SDK, LLM APIs such as OpenAI API, streaming, tool calling, structured outputs, as well as prompt and context design. Understanding of model limitations, token consumption, latency, and cost.

• Agents & MCP. Cloudflare Agents SDK, Project Think, and Model Context Protocol (MCP). Understanding of stateful agents, orchestration, tool binding, and human approvals.

• Cloud & Runtime. Cloudflare Workers, Durable Objects, D1, and R2. Experience with serverless applications, persistent state, and separation of application data and file storage.

• Backend & Interfaces. Hono, HTTP/REST APIs, and Zod for runtime validation. Secure authentication, role-based access control, and robust integration of external systems.

• Frontend. React, Vite, Tailwind CSS, shadcn/ui, and Vercel AI Elements. Development of intuitive chat and agent interfaces with streaming, status indicators, and approval workflows.

• Data & Persistence. SQL/SQLite, Prisma ORM, and database migrations. Clean data models as well as strict separation of user and tenant data.

• Development & Quality. Git, pnpm, Vitest, ESLint, Prettier, Wrangler, and CI/CD. Automated testing, systematic error analysis, and traceable changes.

You should be able to work productively across key parts of this stack. While specialized individual components can be picked up as needed, we require solid experience with TypeScript, React, LLM integrations, and the Cloudflare platform.

Additionally, we expect an understanding of embeddings, semantic search, and RAG, as well as typical AI application risks such as hallucinations, prompt injection, and unauthorized tool access.

How do you work?

You think beyond individual frameworks and can thoroughly evaluate technical alternatives. At the same time, you respect existing architecture decisions and develop solutions within shared standards.

You prefer simple, maintainable code, leverage proven components, and avoid reinventing the wheel. You test assumptions early, deliver in manageable iterations, and openly communicate technical limitations.

You deliberately utilize AI-assisted development tools. You verify their output and take responsibility for code quality, security, and functionality. Data privacy, clear permissions, traceable actions, and human approvals are integral parts of your implementation approach.

Our hiring process consists of 4 steps:

1. HR Call

2. Coding Challenge

3. Challenge Review / Technical Interview

4. Meet with Head of Enterprise IT

During the HR Call, we will cover everything that might be interesting for you, but also details on benefits, plans, and administrative things.

You can learn more about our perks on our website.

A 3D office tour is available here.

We’re looking forward to your application!