Reliable AI Deployment on Heterogeneous Host-Accelerator Systems with Programmable RISC-V Accelerators by Exploiting Hardware Reliability Features

Scientific-Disciplinary Group

09/IINF-01 - Electronics

Description

This Incarico di Ricerca position, aligned with the ARCHYTAS project, focuses on methodologies and software tools for reliable AI deployment on a heterogeneous system with a host and two programmable accelerators: a RISC-V scalar multi-core accelerator and a RISC-V vector accelerator. The work will extend deployment frameworks and toolflows, including Deeploy, to support application mapping, code generation, and execution orchestration, while integrating reliability-aware strategies based on hardware features such as error detection/correction, integrity checks, status monitoring, and diagnostics. Research includes deployment and partitioning strategies, toolflow and runtime extensions, runtime monitoring and mitigation, evaluation metrics, fault injection, and experimental validation on representative AI inference case studies. The goal is to advance portable, efficient, and reliable AI deployment on programmable RISC-V heterogeneous systems.

Job posting website

https://bandi.unibo.it/ricerca/incarichi-di-ricerca

Funding body

ALMA MATER STUDIORUM - UNIVERSITA' DI BOLOGNA - - DIPARTIMENTO DI INGEGNERIA DELL'ENERGIA ELETTRICA E DELL'INFORMAZIONE "GUGLIELMO MARCONI"

How to apply

Other

Selection process