Design and prototyping of new Machine Learning algorithms for Predictive Maintenance purposes
Thesis in external company
keywords ARTIFICIAL INTELLIGENCE, MACHINE LEARNING, PREDICTIVE MAINTENANCE
Reference persons EDOARDO PATTI
External reference persons Davide Mazzucchi (email@example.com)
Research Groups DAUIN - GR-06 - ELECTRONIC DESIGN AUTOMATION - EDA, ELECTRONIC DESIGN AUTOMATION - EDA, GR-06 - ELECTRONIC DESIGN AUTOMATION - EDA, ICT4SS - ICT FOR SMART SOCIETIES
Description Zirak is an Italian SME company, established in 2000, with an Italian headquarter and a Portuguese R&D lab, providing high quality Information Technology services and extensive support on automotive and IoT related topics, from architectural analysis and requirements management to implementation and validation of customized state-of-the-art solutions. Zirak has also contributed to several M2M projects (e.g. dynamometric key interfacing system, sanitification system for hospital unit spaces and electric bike-sharing platforms) designing an Internet-of-Things (IoT) platform for predictive logistics/maintenance in collaboration with major European R&D centers, actively contributing to predictive
maintenance projects in the context of Industry 4.0 programs.
The proposed thesis belongs to the IoT sector, which sees as its objectives:
1. the in-depth review of the current machine learning algorithm used by Zirak current predictive logistics platform (REDtag and PREMA), and the proposal of alternatives, based on state of the art technologies, with an eye to handling different type of data so as to satisfy the predictive maintenance use cases that will be identified with the company tutor (e.g. in the automotive industry) also based on data availability of the same data sets;
2. The best suited algorithm shall then be identified and prototyped, with training and application examples on new data sets;
3. Integration of the developed prototypes into our developed backend solution. Functionality testing and polishing;
4. Write documentation of the conducted review on the state of the art, the explored datasets and of the developed prototypes and the process of their integration into the backend solutions.
See also www.zirak.it
Required skills The candidate is required to have the following skills and competences:
Good knowledge of Python.
Good knowledge of Artificial Intelligence methods and Machine Learning techniques.
Good attitude to teamwork.
Good level of spoken and written English.
Basic knowledge of Jupyter framework.
Basic knowledge of the following libraries: Scikit-learn, TensorFlow, Keras or similar.
Basic knowledge of Matlab and Spark framework.
Notes availability to travel to Mondovμ part-time (60% time).
Deadline 10/06/2023 PROPONI LA TUA CANDIDATURA