Agentic Software Engineer
Zühlke Group23 hours ago
Belgrade, SerbiaSenior
Responsibilities
- Design and build multi-agent workflows with durable state, clear hand-offs, explicit failure handling, stop conditions, retry budgets, and model-selection logic.
- Engineer agent loops, harnesses, context and tool access through MCP, instruction files, skills, custom agents, hooks, and CI checks.
- Implement spec-driven development with traceability from requirements to code and tests, including independent evaluators and hidden acceptance tests.
- Build quality gates using deterministic checks, evaluations, AI judgment, and human decisions supported by evidence.
- Harden agents against prompt injection, run them in isolated least-privilege sandboxes, and make agent behavior observable through audit trails, cost metrics, and lead-time metrics.
- Work with client teams to deploy agents, harnesses, and workflows to production, and remain accountable for adoption and delivered value after go-live.
- Apply responsible AI and governance practices, including human oversight, logging, documentation, and traceability.
Requirements
- At least seven years of professional software engineering experience and a record of owning code quality in production systems.
- A working agentic system built personally at work or in free time, with the ability to demonstrate it as a repository or live system during selection.
- Hands-on experience with multi-agent orchestration, workflows, hand-offs, state management, failure handling, context engineering, and the Model Context Protocol.
- Experience with agent frameworks such as Claude Agent SDK, Azure AI Foundry, or LangGraph, and with evaluation gates, prompt-injection guardrails, and agent tracing tools such as Langfuse.
- Experience engineering AI coding-agent harnesses using instruction files, skills, custom agents, and hooks, plus agent loops with independent evaluators, stop conditions, and cost limits.
- Strong programming skills in Python or TypeScript; Java or .NET experience is a plus.
- Solid CI/CD and Git practice; GitLab, containers, and Kubernetes experience is a plus.
- Clean, tested, well-structured coding practices and experience taking software into production with clients.
- Working knowledge of responsible AI and governance, including EU AI Act obligations related to human oversight, logging, and documentation.
- Delivery experience in regulated finance or health industries is a strong plus, along with confident English communication and effective collaboration with international teams and non-engineers.
Tech Stack
Categories
About Zühlke Group
Zühlke Group provides technology consulting, product engineering, and digital solutions for enterprises, especially in regulated sectors such as healthcare and finance. It builds custom software, connected devices, and data-driven services, and supports clients from strategy through implementation and operations. Founded in Switzerland in 1968 and headquartered in Schlieren, Zurich, the firm is partner-owned, operates across Europe and Asia, and also invests in startups via Zühlke Ventures.