Pierre Mecca

• Managing and supporting EMEA-based customers in the adoption and implementation of edge AI solutions built on Hailo products. • Providing in-depth technical expertise and pre/post-sales support, including system design, debugging, and performance optimization. • Leading Hailo's software enablement efforts for the ASTRIAL platform, a joint solution developed by EBV Elektronik and System Electronics, including application • Conducting technical presentations and workshops to demonstrate the capabilities and benefits of Hailo products to customers. • Took ownership of strategic software enablement initiatives beyond the tradition FAE scope, including the porting and optimization of the Whisper speech recognition models for Hailo devices.

Responsible of Digital Mixed Signal Verification for Automotive BU products • Development and automation of UVM-based environments for DMS simulations • Definition of verification plans and analysis of Functional Coverage and Code Coverage • Planned and managed project software releases.

• Design and verification of digital modules for mixed-signal integrated circuits. • Implemented Python-based methodologies for verification.

The activity consisted in the deployment of deep learning algorithms on embedded devices, working both at framework level (adapting the neural network model operators to a format accepted by vendor converters) and at execution level (using C/C++ APIs to run the model on the device). Tools: Tensorflow, ONNX, PyTorch, SNPE, QTCreator, Git Platforms: Qualcomm , Renesas , Texas Instruments, Others Languages: C/C++, Python

The internship was mainly focused on the deployment of deep neural networks on embedded platforms for autonomous driving, from neural networks surgery at framework level (TensorFlow) to the optimized execution on the target hardware. Secondary activities included the development of firmware for automotive camera sensor driver and the management of video streams Tools: TensorFlow, SNPE, Android SDK Programming Languages: C++, Python

Thesis subject: "Advanced Hardware Solutions for Neural Networks Inference in Autonomous Vehicles" The activity was focused on the evaluation of embedded platforms for neural networks inference, as well as on optimization techniques to deploy AI models on hardware platforms with limited parallelism like FPGAs. The project covered a large spectrum of topics, from neural networks definition and optimization using Caffe framework to implementation of basic operators on FPGA using Xilinx Vivado. Programming Languages: C++, Python, VHDL

Characterization of an NG-PON2 optical receiver