Agentic AI Engineer

Project overview

For our client, we are looking for an Agentic AI Engineer (m/f/d) to design and develop production-ready AI agents and the necessary automation and integration pipelines. The focus is on autonomous or semi-autonomous agent systems that can plan tasks, utilize deterministic tools and enterprise knowledge, and execute complex development steps as independently as possible.

The role combines RAG and Knowledge Base architectures with Context Engineering, secure Tool Calling, Sandboxing, and cost-conscious system design. Additionally, evaluation, observability, MCP, vector databases, and suitable guardrails are relevant knowledge for robust agent systems. These additions cover typical near-production Agentic AI topics such as Tool Calling, Evaluation, Tracing, and MCP.

Responsibilities

Must-have Skills

Nice-to-have Skills

• Conception and development of automation pipelines for autonomous AI agents
• Development of agentic systems that can handle complex tasks and development steps as independently as possible
• Conception and implementation of RAG systems
• Setup and connection of Knowledge Bases
• Development of deterministic tools and functions for controlled use by AI agents
• Context Engineering and control of information provided to agents
• Cost-optimized architecture design and monitoring of resource/token consumption
• Implementation of secure sandboxing mechanisms for controlled execution of agentic actions
• Integration of agents with APIs, data sources, and external systems
• Setup of error handling, retry mechanisms, and controlled execution paths. These mechanisms are part of typical near-production agent workflows.
• Setup of evaluation and observability mechanisms for measuring agent behavior, quality, costs, and latency.
• Profound practical experience with Agentic AI / AI Agents
• Development of RAG systems
• Setup and technical integration of Knowledge Bases
• Development of deterministic tools / Function Calling
• Context Engineering
• Cost-aware design, especially regarding token consumption, model selection, and execution costs
• Sandboxing and security for controlled agent execution
• Very good software engineering skills for developing automation and agent pipelines
• Python and/or TypeScript (researched addition) as typical implementation languages for near-production agent systems.
• Experience with vector databases, embeddings, chunking, retrieval, and reranking. These components are particularly relevant for RAG and knowledge systems.
• Knowledge of Model Context Protocol (MCP) for standardized tool and system integration.
• Experience with agent frameworks like LangGraph, CrewAI, LlamaIndex, or OpenAI Agents SDK.
• Experience with LLMOps, evaluation, and regression testing for AI systems.
• Tracing and observability of agents, tool calls, as well as costs and latencies.
• Knowledge of Human-in-the-Loop / Approval Gates, Guardrails, and secure tool execution.
• Experience with API design and the integration of external systems and databases.

Project details

01.01.2027 or ASAP

Contract Type: Service

Duration / End: 12 months with option for extension

Koblenz and Remote

Utilization/Volume: Full-time

• Start: 2027-01
• Duration: 12 Monate
• Remote: hybrid
• Location: Koblenz, Deutschland, DE
• Languages: German