PhD in Efficient and Reliable AI Inference, AI Accelerator Optimization, and Energy-Efficient Deep Learning University of Trento in Italy

The University of Trento and the HiCREST Laboratory invite applications for a PhD student position in efficient and reliable AI inference. The project is focused on AI accelerator design and use, inference optimization, and energy-efficient deep learning, with particular attention to transformer-based models.

Research topics include quantization, sparsification, and pruning to reduce computational cost, memory footprint, and energy consumption while studying the effects on accuracy, runtime performance, and reliability. The work aims to develop tools, benchmarks, and optimization methods for reproducible AI systems research and may contribute to the Archytas framework.

Eligible backgrounds include Computer Science, Computer Engineering, Artificial Intelligence, Data Science, Electrical Engineering, or related fields. Useful skills include Python, C/C++, PyTorch, TensorFlow, ONNX Runtime, CUDA/GPU programming, benchmarking, HPC, model compression, and reliability analysis.

The position is fully funded for 3 years with a stipend of about €1.7K/month after tax. There is also an internship opportunity at Covision Lab. The application is an expression of interest: send your CV and a 1-page research/motivation statement to [email protected].

Deadline: 2026-07-31.