Required skills
- Python
- Pydantic
- AsyncIO
- Clean Code
- Design Patterns
- Pytest
- Mocking
- Ruff
- Git
- GitLab CI/CD
- SemVer
- Docker
- LangGraph
- State Machines
- Human-in-the-loop
- Prompt Engineering
- Model Context Protocol (MCP)
- FastMCP
- Document Processing
- Retrieval-Augmented Generation (RAG)
- Vector Embeddings
- Semantic Search
- Hybrid Search
Languages: German
Job description
The role involves conceptualizing, implementing, and testing production-ready agent-as-code solutions using Python and LangGraph. The AI Engineer will integrate tools, realize RAG and document processing pipelines, and prepare agents for containerized deployments.
Project details
Jan 2027
End of 2027+
remote and possibly Frankfurt
Scope: Full-time
Note: Several consultants are sought
Start 18.01.2027; Pitch *;
The role conceptualizes, implements, and tests production-ready agent-as-code solutions in the use case pods.
Based on Python and LangGraph, the role integrates tools via MCP, realizes RAG and document processing pipelines with central platform services, and prepares well-structured agents for containerized deployments. For machine learning use cases, the AI Engineer and Data Scientist implement the production-ready solution in Dataiku.
Required Skills (Must-Have)
-
Software Engineering:
Excellent knowledge of Python (comprehensive practice with modern type hints, Pydantic, AsyncIO, modular packaging).
Mastery of professional software engineering practices: Clean Code, Design Patterns, Module and Integration Tests (Pytest, Mocking), Linting & Formatting (Ruff).
Proficient use of Git, GitLab CI/CD, SemVer, and containerization (Docker).
Agentic AI & Orchestration:
Solid practical experience with LangGraph (state graphs, state machines, check-pointing, human-in-the-loop / approval workflows).
Deep understanding of agent architectures, tool calling, structured outputs (JSON Schema Validation), and prompt engineering methods.
Hands-on experience with the Model Context Protocol (MCP), especially with client and server implementations via FastMCP.
RAG, Document Processing & Retrieval:
Solid experience in extracting and processing unstructured document formats (PDF, DOCX, PPTX, HTML, XML, XLSX) using modern parsing tools and libraries (e.g., Docling, Unstructured, PyPDF/pdfplumber, OCR methods).
Conceptualization and implementation of robust, structure- and semantics-aware chunking methods (recursive, Markdown-based) for optimal retrieval quality.
Practical experience with Retrieval-Augmented Generation (RAG): vector embeddings, semantic and hybrid search, re-ranking concepts.
Practical experience in using vector data
- Start: 2027-01
- Duration: 12 Monate
- Remote: hybrid
- Location: remote, Deutschland, DE
- Languages: German
Indexed by 9am's meta-search. You are a job provider? Claim or manage this listing
Found on FreelancerMap DE