Academic Internship: Optimized ML inference on cloud compute instances
Bosch Car Multimedia Portugal
Braga, Braga, pt
há 7 horas

Job Description

Cloud platforms like AWS or Azure became a very attractive alternative for companies to outsource their infrastructure, especially when it comes to high performance computing and scalable systems.

This allows companies to perform very compute intensive operations, such as the training of AI models, on the spot, paying the compute resource as a service.

Since those services are paid based on actual compute time, the time an AI model needs to be trained or inferred directly translates to costs.

Thus, having an optimized ML framework and selected the right compute instances for optimal GPU utilization becomes a crucial topic.

Main tasks :

  • The candidate should be capable of analyzing the current state of the art literature and summarize the important points;
  • Get an overview of different compute options on public clouds, e.g. AWS Inferentia;
  • Understand the Keras based ML framework to be used as part of the Proof of Concept;
  • Formulation of the problem statement and designing the solution architecture to solve it;
  • Implementation of selected approach using the provided Keras based ML framework and available GPU compute instances on AWS;
  • Prepare a test setup to evaluate and compare different experiments, including selected logs and metrics regarding GPU utilization;
  • Perform different experiments and gather results to show effectiveness of selected approach;
  • Write documentation of implemented approach (preferred in English);
  • Qualifications

    What distinguishes you :

  • Attending Master degree in Software Engineering, Computer Science or similar;
  • Basic SW development skills in Python & Containerization (e.g. Docker);
  • Basic knowledge in Machine Learning frameworks (e.g. TensorFlow);
  • Basic statistics knowledge (is a plus);
  • English proficiency comparable to skill level B2 or higher.
  • Reportar esta oferta de trabalho

    Thank you for reporting this job!

    Your feedback will help us improve the quality of our services.

    Meu email
    Ao clicar em "Continuar", autorizo a neuvoo a processar os meus dados e a enviar-me alertas de e-mail, conforme detalhado na Política de Privacidade da neuvoo . Posso retirar o meu consentimento ou cancelar a subscrição a qualquer momento.
    Formulário de candidatura