ai engineer

Orbitant provides engineering expertise to help companies accelerate their development through strategic technical decision-making and the implementation of scalable, AI-driven solutions.

• Design, build, and ship production-grade AI agents and LLM-based systems for national and international clients;
• Manage projects from pilot to production, including defining evaluations, ensuring traceability, and meeting business outcomes;
• Implement and maintain RAG systems, including chunking strategies, retrieval, reranking, and ingestion pipelines;
• Design multi-step agent and workflow architectures, including state management, tool use, error recovery, and cost management;
• Build offline and online evaluations using tools like Langfuse to benchmark against real outcomes;
• Design and integrate MCP servers to connect agents with existing client systems such as ERPs and CRMs;
• Contribute to the internal AI-first platform by developing reusable skills, libraries, and templates;
• Collaborate with business stakeholders to define project outcomes before determining technical architecture;
• Review and improve code, maintain high quality standards, and work with legacy code when necessary;
• Mentor junior teammates and share knowledge through technical sessions and documentation;
• Propose new techniques and tools that provide genuine value;
• Maintain accountability for production systems regarding cost, latency, observability, and reliability.

• Experience building production systems with LLMs, including agents, RAG, and generation pipelines;
• Proficiency in Python as a primary working language;
• Hands-on knowledge of agent or workflow orchestration frameworks like LangGraph or LangChain;
• Experience implementing evaluations and traceability for LLM-based systems using tools like Langfuse;
• Understanding of prompt engineering techniques such as few-shot, CoT, and structured output;
• Experience with Vector DBs like pgvector, Qdrant, Pinecone, Weaviate, or Chroma;
• Ability to evaluate trade-offs between different providers and models regarding cost, latency, and capabilities;
• Mastery of the Anthropic ecosystem, including Claude Code and Claude SDK;
• Experience designing or consuming MCP servers;
• Strong engineering practices including clean code, testing, CI/CD, and version control;
• English level sufficient for communication with an international team and clients;
• Nice to have: TypeScript/Node, Cloud infrastructure (AWS/GCP/Azure), Docker/Kubernetes, fine-tuning or lightweight training (LoRA), experience mentoring others, knowledge of AI security and privacy, OSS contributions or public AI side-projects.

• Salary between €40,000 and €60,000 gross per year;
• Access to a top office in Madrid with free attendance;
• Latest generation MacBook;
• Claude license;
• Training and mentoring programs;
• Flexible benefits including the possibility of health insurance;
• Referral bonus.