Lead AI Agent Architect for HPC Systems

Barcelona Supercomputing Center; Barcelona; On-site; EUR 70,000 - 110,000; Full time

Benefits offered by this job

Restaurant tickets

Private health insurance

Relocation assistance

Flexible working hours

Job summary

The Barcelona Supercomputing Center (BSC-CNS) in Barcelona invites applications for a researcher to design and implement innovative language-agent techniques in a high-performance computing setting. You will lead the development of agentic workflows, integrate MCP and function calling into core models, and supervise robust evaluation frameworks.

You will collaborate with the infrastructure team, mentor junior researchers, publish in top conferences, and contribute to Open Source projects.

Qualifications

• PhD or Master in Computer Science, Artificial Intelligence or related field.
• Proven track record in developing LLM-based agents or complex AI pipelines.
• Expert knowledge of Python and DL frameworks like PyTorch or JAX.
• Deep understanding of Transformer architectures and agentic AI.
• Hands-on with post-training frameworks TRL, FastChat, NeMo-RL.
• Experience with distributed training and large-scale deployment.
• Excellent written and spoken English.
• Strong publication record in top conferences and OSS contributions.
• Spanish is highly valued; Catalan/Basque/Galician advantageous.

Responsibilities

• Lead design and implementation of agentic workflows with multi-step reasoning, tool-use, and autonomous decision-making.
• Architect integration of MCP and Function Calling into core language models.
• Supervise development of agent evaluation frameworks for reliability, safety, and tool interaction.
• Collaborate with infrastructure teams for scalable deployment in HPC environments.
• Provide technical leadership and mentorship to junior researchers and define roadmaps.

Skills

LLM agents

Function calling

AI pipelines

Python

Deep learning

PyTorch

JAX

Transformer models

English

Open source

Education

PhD or Master in CS/AI

Tools

TRL

FastChat

NeMo-RL

PyTorch

JAX

Job description

The Barcelona Supercomputing Center (BSC-CNS) in Barcelona invites applications for a researcher to design and implement innovative language-agent techniques in a high-performance computing setting. You will lead the development of agentic workflows, integrate MCP and function calling into core models, and supervise robust evaluation frameworks.

You will collaborate with the infrastructure team, mentor junior researchers, publish in top conferences, and contribute to Open Source projects.