Postdoc: Dynamic Parallelization of Sparse Codes for ML

HiPEAC; Lyon; Sur place; EUR 42 000 - 54 000; Plein temps

Résumé du poste

Inria Lyon seeks a Post-Doctoral Researcher for 14 months to optimize dynamic parallelization of sparse codes in high-performance computing and machine learning, focusing on compiler algorithms and scheduling.

The fellow will propose code optimizations and data structures for scaling sparse propagation, address runtime scheduling by relying on existing parallel runtimes, and validate the approach on benchmarks from standard collections.

This position is funded by the French PEPR NumPEx program.

Qualifications

• PhD in CS/Applied Math required.
• Experience with compiler/runtime algorithm development.
• Background in HPC and machine learning is desirable.

Responsabilités

• Propose code optimizations and data structures for scaling sparse propagation.
• Address runtime scheduling by relying on existing parallel runtimes.
• Validate the complete approach on benchmarks using sparse tensors from standard collections.

Connaissances

Compiler optimization

Runtime scheduling

Parallel computing

Machine learning

Formation

PhD in Computer Science / Applied Math

Outils

BLAS/LAPACK

HPC runtimes

Description du poste

Inria Lyon seeks a Post-Doctoral Researcher for 14 months to optimize dynamic parallelization of sparse codes in high-performance computing and machine learning, focusing on compiler algorithms and scheduling.

The fellow will propose code optimizations and data structures for scaling sparse propagation, address runtime scheduling by relying on existing parallel runtimes, and validate the approach on benchmarks from standard collections.

This position is funded by the French PEPR NumPEx program.