Machine Learning / AI Systems Engineer

Madrid (preferably) / Remote

Role

As an AI Systems Engineer at Aqmen, you will play a central role in building the agentic AI foundation that powers Aqmen’s products. You will design and implement LLM-driven agents, multi-step reasoning workflows, and tool-augmented pipelines that guarantee correctness, auditability, and reproducible behaviors in high-stakes environments.

You will work across the entire lifecycle of our agent stack: prompt and context engineering, finetuning, tool and API orchestration, agent graph design (e.g., LangGraph), evaluation harnesses, and production deployment. Your work will directly shape how LLMs interact with structured data, Excel models, verification tools, and the domain-specific logic central to our products.

This role is ideal for an engineer who is excited about LLM behavior and building trustworthy AI systems, and wants to help define a new standard for safe, deterministic agentic AI in the workplace. An analytical mindset, strong problem-solving instincts, and a natural inclination toward structured reasoning are essential.

Key Responsibilities

Agentic AI & LLM Engineering

  • Design, implement, and iterate on LLM-driven agents, including tool-calling workflows, graph-based agent controllers, and multi-step reasoning pipelines.

  • Develop and maintain structured prompting and instruction-following systems that yield deterministic and verifiable outputs.

  • Build evaluation frameworks for LLM agents, including behavioral tests, reproducibility checks, and safety/consistency validation.

  • Integrate LLMs with internal tools, Python functions, and external APIs to enable complex multi-step workflows.

  • Apply analytical problem-solving techniques to diagnose agent failure modes, improve reasoning reliability, and reduce ambiguity and hallucination.

Infrastructure, Security, Collaboration, & Production Deployment

  • Deploy agent workflows into production using LangGraph and containerized services.

  • Monitor agent behavior in production, including error rates, tool-call reliability, hallucination detection, and output correctness.

  • Build scalable pipelines for prompt versioning, dataset creation, fine-tuning, and structured logs for audits.

  • Build agent workflows that are traceable, auditable, and compliant, including structured logs for every model decision and tool invocation.

  • Implement guardrails for sensitive data handling in LLM contexts, including data minimization, pseudonymization, and access-controlled tool use.

  • Collaborate with the product and design teams to translate complex AI workflows into clean UI components.

  • Design clean APIs and backend contracts that support deterministic, agent-driven operations.

 Requirements

  • 3–5+ years as a machine learning engineer, software developer, or related roles.

  • Strong Python proficiency with production systems (FastAPI, Pydantic, Pandas).

  • Familiarity with agentic AI patterns, tool calling, structured outputs, and workflow orchestration.

  • Strong software engineering fundamentals and ability to design and test reliable, deterministic systems. Proficiency in object oriented programming. 

  • Strong analytical and quantitative reasoning skills; ability to decompose complex problems into simple, granular agent steps.

Nice to Have

  • Experience building or maintaining front-end features using React with Javascript/Typescript.

  • Experience working with LangChain/LangGraph and backend services that interface with LLM agents, including FastAPI and serverless endpoints.

  • Comfort working with database schemas, query optimization, and integrating agent workflows with databases such as PostgreSQL, ClickHouse, or MongoDB.




Apply