Posizione
Florence office or remote
Duration: 3 to 4 months starting June 2024
Start Date: 3rd of June
Internship Overview:
As a Neural Network Model Optimization Intern, you will join our dynamic team to enhance the capability of our proprietary accelerators by building a specialized ModelZoo. This project will involve acquiring publicly available neural network models and meticulously optimizing and debugging them for compatibility with the Axelera Software Development Kit (SDK).
Key Responsibilities:
Model Acquisition: Identify and procure relevant neural network models from public repositories based on Axelera’s market roadmap.
Optimization: Utilize the Axelera SDK to manually optimize these models for enhanced performance and efficiency on Axelera accelerators.
Debugging and Testing: Conduct thorough debugging to resolve any issues encountered during the optimization process. Implement rigorous testing routines to ensure models are robust and reliable.
Documentation: Maintain detailed documentation of optimization strategies, modifications, and performance benchmarks.
Collaboration: Work closely with software developers and hardware engineers to align model optimization strategies with hardware capabilities and project goals.
Innovation: Provide innovative ideas and suggestions to improve model performance and integration processes.
Benefits:
Opportunity to work with state-of-the-art AI technologies.
Mentorship from industry experts in AI and accelerator technologies.
Potential for full-time employment upon successful completion of the internship.
Caratteristiche del candidato
Requisiti tecnici e conoscenze informaticheKnowledge in neural networks, machine learning, and artificial intelligence.
Prior exposure to machine learning frameworks (e.g., TensorFlow, PyTorch).
Prior exposure to programming languages such as Python or C++. Lingue straniereE' richiesta la conoscenza delle seguenti lingue
1. Inglese: buono (B2-C1)
Titoli preferenzialia degree in Computer Science, Electrical Engineering, or a related field
Area disciplinarescientifico,ingegneriaDisponibilità al trasferimentoNon richiestaDisponibilità trasferimento all'esteroNon richiesta