Senior AI Agent Engineer: Build & Validate MultiStep Agents

FirstIgnite; Valencia; Presencial; EUR 60.000 - 90.000; Jornada completa

Descripción de la vacante

FirstIgnite in Valencia is building AI agents for document-heavy workflows. As a Senior AI Agent Engineer you will design and ship long-running, multi-step agents that call APIs and process long documents, ensuring accuracy and traceability before human handoff.

You will work with product and engineering teams to turn vague quality goals into measurable targets, compare models and prompts, and build eval suites. Prior experience with LLMs, agent frameworks, and API integrations is essential.

Formación

• 3+ years of engineering experience with LLM/agent systems.
• Experience evaluating agents, not just models.
• Experience integrating APIs into agent calls.
• Experience with document pipelines and truth sources.
• Familiarity with at least one LLM evaluation framework.
• Knowledge of multi-step, tool-using agent design.

Responsabilidades

• Design and ship long-running, multi-step agents on various AI SDKs and tooling.
• Wrap APIs as tools an agent can call over MCP.
• Get structured data out of long documents and match against records.
• Stand up eval suites using evaluation frameworks and tooling.
• Build citations and confidence signals for fast human review.
• Collaborate with product and domain experts to define measurable quality goals.
• Instrument production traffic and turn interactions into datasets.
• Compare models and prompts across options and assess costs in latency and quality.
• Bootstrap quality signals with synthetic data and edge cases.
• Write the templates, docs, and tooling for the team.

Conocimientos

LLM systems

Agent frameworks

API integration

Document processing

Evaluation frameworks

Herramientas

OpenAI Agents SDK

Anthropic Agent SDK

Vercel AI SDK

LangGraph

MCP

Temporal Cloud

Descripción del empleo

FirstIgnite in Valencia is building AI agents for document-heavy workflows. As a Senior AI Agent Engineer you will design and ship long-running, multi-step agents that call APIs and process long documents, ensuring accuracy and traceability before human handoff.

You will work with product and engineering teams to turn vague quality goals into measurable targets, compare models and prompts, and build eval suites. Prior experience with LLMs, agent frameworks, and API integrations is essential.