Runtime Architect for Co-Designed AI Accelerators

HiPEAC; Bordeaux; Sur place; EUR 45 000 - 70 000; Plein temps

Résumé du poste

Inria Bordeaux seeks an engineer to develop and optimize a runtime system for machine learning chains, focusing on performance and energy efficiency through close collaboration with hardware and software teams.

The role will advance a co-designed runtime that integrates with CAMELIA's software stack, validating software in CI, documenting the work, and ensuring interoperability across components and configurations.

Qualifications

• Excellent programming in C and C++.
• Linux environments with Git and CI.
• Parallel, distributed and GPU programming on HPC.
• Machine learning chains for training/inference.
• Experience in validation and verification of code.

Responsabilités

• Develop and optimize the runtime system.
• Define, tune and validate scheduling policies.
• Validate software in CI processes.
• Prepare, validate and maintain software documentation and interoperability.

Connaissances

C/C++ programming

Linux development

Parallel and GPU programming

ML pipelines

Code validation & verification

Outils

Git

CI frameworks

Linux

HPC platforms

Description du poste

Inria Bordeaux seeks an engineer to develop and optimize a runtime system for machine learning chains, focusing on performance and energy efficiency through close collaboration with hardware and software teams.

The role will advance a co-designed runtime that integrates with CAMELIA's software stack, validating software in CI, documenting the work, and ensuring interoperability across components and configurations.