Solutions Architect: LLM Evaluation & Agentic Pipelines

NVIDIA; France; Sur place; EUR 67 000 - 116 000; Plein temps

Résumé du poste

NVIDIA's Worldwide Field Operations team seeks an AI-focused Solution Architect to evaluate neural networks and agentic pipelines. You will interact with customers to understand needs, demonstrate proof-of-concept solutions, and drive adoption of NVIDIA tools across EMEA.

The role emphasizes benchmarking systems, model selection, and fine-tuning within end-to-end agentic architectures. Ideal candidates have advanced degrees and 5+ years in building or evaluating deep neural networks, with strong

Qualifications

• MS/PhD or equivalent in CS, data science, engineering, physics or maths.
• 5+ years building or evaluating deep neural networks.
• Experience in ML/DL/DL data science with emphasis on deep learning.
• Knowledge of LLMs or VLMs and retrieval systems.
• Familiarity with evaluation libraries like Eval Harness or NVIDIA Evaluation Microservice.
• Understanding of modern evaluation benchmarks and extending them.

Responsabilités

• Collaborate with customers to understand technology and propose best solutions.
• Promote adoption of NVIDIA open-source evaluation tools and establish standard methods.
• Develop solutions based on NVIDIA and open-source LLM tech for agentic pipelines.
• Work with engineering, product and sales to design customer-centric solutions.
• Build industry expertise and contribute to integrating NVIDIA tech into architectures.

Connaissances

Deep neural networks evaluation

LLMs & retrieval systems

Effective communication

Formation

MS/PhD or equivalent

Outils

Eval Harness

NVIDIA Evaluation Microservice

NeMo

Description du poste

NVIDIA's Worldwide Field Operations team seeks an AI-focused Solution Architect to evaluate neural networks and agentic pipelines. You will interact with customers to understand needs, demonstrate proof-of-concept solutions, and drive adoption of NVIDIA tools across EMEA.

The role emphasizes benchmarking systems, model selection, and fine-tuning within end-to-end agentic architectures. Ideal candidates have advanced degrees and 5+ years in building or evaluating deep neural networks, with strong