Software Engineer Agentic AI (Coding Agent Harness)
- Employment type
- Full-time
- Workload
- 100%
- Location
- Bellinzona
- Languages
- German, German
- First posted
Are you already spending more time with prompts, skills, and evaluations than with boilerplate code? Do you want to design the coding agent that the developers at the Post work with daily, and influence what it knows, what it is allowed to do, and how it improves? Then you might be exactly the right person for us. The IT of die Schweizerische Post AG is changing the way software is developed: AI agents are taking over a growing part of the work, while humans set the goals, check the results, and control the processes. Our developers are already working with coding agents today. With you, we will turn this into a coding agent for the Post that a company can rely on.
What you can achieve
- You specify, create, test, and deploy our coding agent, which is pre-configured with skills, prompt templates, rules, and plug-ins.
- Additionally, you maintain the basic configuration and our skill library so that the teams receive a version they can rely on: tested, versioned, and easy to undo.
- You write evaluations for skills, prompts, and agent workflows so that with every change it is clear whether it actually improves the agent or merely changes it.
- At the same time, you integrate security mechanisms directly into the agent, for example permissions, sandboxing, and hooks, so that the secure way is also the easiest way.
- You connect our internal systems via MCP and evaluate new Large Language Models based on the actual tasks of our developers.
- In addition, you work together with the pilot teams and the growing number of engineers who use the agent daily, and translate the insights gained into better skills and standards.
What you bring with you
- You have several years of practical experience in software development, for example with Python, TypeScript, Go, or Java, including APIs and CI/CD pipelines such as GitHub Actions.
- Additionally, you bring practical experience in building solutions with Large Language Models (LLMs) and coding agents that go beyond pure chat use, for example in the form of skills, prompts, tools, MCP servers, or agent workflows that are actually used productively.
- You have experience in using GenAI tools pragmatically in real software development, and have a clear understanding of where they create added value and where human judgment is still required.
- Furthermore, you have a strong understanding of security for agents, including secret management, dependencies, permissions, input and output controls, and data protection.
- You have very good oral and written English skills. Knowledge of a national language is an advantage.-
You do not have to meet every single criterion. If you bring a strong technical foundation and have already developed something concrete with AI agents, we would be happy to hear from you. It is more important to us how you think and how you learn than whether your profile matches all the keywords in the job description perfectly. Help us make our company the "Post of tomorrow"!
Automatically translated from the original.
Posted 3 days ago