AI Agent & Tool-Use Engineer for LLMs in HPC

Barcelona Supercomputing Center; Barcelona; Presencial; EUR 80.000 - 110.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Restaurant tickets

Private health insurance

Relocation support

Flexible working hours

Training plan

Descripción de la vacante

Barcelona Supercomputing Center is seeking an AI Agent & Tool-Use Engineer (RE2) to join the Language Modeling Team. You will design and test agentic workflows, enhance model interactivity with MCP, and run post-training experiments in a high-end HPC environment.

The role requires hands-on experience with LLM fine-tuning, PyTorch, and tool-use patterns like ReAct. English proficiency is essential; Spanish language skills are valued and relocation support is provided.

Formación

• Degree in Computer Science, Mathematics or related field.
• Solid experience in Fine-tuning and Post-training of Large Language Models.
• Proficiency in Python and Deep Learning libraries like PyTorch.
• Experience with Agentic AI patterns such as ReAct or autonomous tool-use.
• Familiarity with TRL, FastChat, NeMo-RL and MCP is desirable.
• Excellent English skills (written and spoken).

Responsabilidades

• Implement and test agentic workflows with multi-step reasoning and tool-use.
• Develop and refine Function Calling interfaces using MCP to enhance model interactivity.
• Run post-training experiments to improve agentic performance of models.
• Build and maintain evaluation pipelines for model-tool interactions.
• Collaborate with data teams to curate instruction datasets for dialogue and function execution.

Conocimientos

Python

PyTorch

ReAct

Post-training

Fine-tuning

English (written and spoken)

Educación

BSc in Computer Science or Mathematics

Herramientas

TRL

FastChat

NeMo-RL

Model Context Protocol (MCP)

Descripción del empleo

Barcelona Supercomputing Center is seeking an AI Agent & Tool-Use Engineer (RE2) to join the Language Modeling Team. You will design and test agentic workflows, enhance model interactivity with MCP, and run post-training experiments in a high-end HPC environment.

The role requires hands-on experience with LLM fine-tuning, PyTorch, and tool-use patterns like ReAct. English proficiency is essential; Spanish language skills are valued and relocation support is provided.