Senior AI Engineer (Agentic AI & LLM Systems)

Your tasks:

We are looking for an experienced AI Engineer to join our growing AI team at Beroe X nnamu. You'll play a key role in developing intelligent, agentic AI systems using cutting-edge large language models (LLMs), multi-agent orchestration, and retrieval-augmented generation (RAG). This is a hands-on role combining software engineering, ML/NLP expertise, and a passion for building next-gen autonomous agents.

You'll collaborate closely with AI leads, backend engineers, data engineers, and product managers to bring scalable and intelligent systems to life, integrated into real-world procurement and business applications.

Key Responsibilities:

• Design and implement agentic AI pipelines using LangGraph, LangChain, CrewAI, or custom frameworks
• Build robust retrieval-augmented generation (RAG) systems with vector databases (e.g., FAISS, Pinecone, OpenSearch)
• Fine-tune, evaluate, and deploy LLMs for task-specific applications
• Integrate external tools and APIs into multi-agent workflows using dynamic tool/function calling (e.g., OpenAI JSON schema)
• Develop memory modules such as short-term context, episodic memory, and long-term vector stores
• Build scalable, cloud-native services using Python, Docker, and Terraform
• Collaborate in agile, cross-functional teams to rapidly prototype and ship ML-based features
• Monitor and evaluate agent performance using tailored metrics (e.g., success rate, hallucination rate)
• Ensure secure, reliable, and maintainable deployment of AI systems in production environments

Your profile:

• 6-8 years of professional experience in machine learning, NLP, or software engineering
• Strong proficiency in Python and experience with ML libraries like PyTorch, TensorFlow, scikit-learn, and XGBoost
• Hands-on experience with LLMs (e.g., GPT, Claude, LLaMA, Mistral) and NLP tooling such as LangChain, HuggingFace, and Transformers
• Experience designing and implementing RAG pipelines with chunking, semantic search, and reranking
• Familiarity with agent frameworks and orchestration techniques (e.g., planning, memory, role assignment)
• Deep understanding of prompt engineering, embeddings, and LLM architecture basics
• Design systems with role-based communication, coordination loops, and hierarchical planning. Optimize agent collaboration strategies for real-world tasks.
• Solid foundation in microservice architectures, CI/CD, and infrastructure-as-code (e.g., Terraform)
• Experience integrating REST/GraphQL APIs into ML workflows

Bonus Qualifications:

• Experience with RLHF, LoRA, or parameter-efficient LLM fine-tuning
• Familiarity with CrewAI, AutoGen, Swarm, or other multi-agent libraries
• Exposure to cognitive architectures like task trees, state machines, or episodic memory
• Prompt debugging and LLM evaluation practices
• Awareness of AI security risks (e.g., prompt injection, data exposure)

Why us?
4-Day Work Week & Up to 24 Days Off. Hybrid options in Berlin or Munich. 8 weeks/year work abroad option.