Smart Safety Improvement and Scaling - ON-523Project type: Research
Desired discipline(s): Engineering - computer / electrical, Engineering, Engineering - other, Computer science, Mathematical Sciences, Mathematics
Company: Vale Canada HQ
Project Length: 4 to 6 months
Preferred start date: 10/25/2021
Language requirement: English
Location(s): Toronto, ON, Canada; Montréal, QC, Canada; Vancouver, BC, Canada; Canada; Canada; Canada; Canada
No. of positions: 1
Desired education level: Master's
About the company:
Vale is a global mining company with its headquaters in Brazil. Vale is a leader in the production of iron ore and the second-largest producer of nickel. Vale’s nickel mining and base metal operations headquater is located in Ontario. Vale’s base metals headquater is in downtown Toronto and the base metal technology development team is located in Mississauga, comprising of over 350 employees. Vale’s mission is to transform natural resources into prosperity and sustainable development. Vale exists to improve lives and transform the future, Together. Vale believes that mining is essential to the world's development. Vale only serves society when it generates prosperity for all and takes care of the planet.
Describe the project.:
The project is an applied data science research focusing on predicting safety incidents.
The main goal of this project is to propose data preparation and modeling techniques to predict safety incidents and compare the proposed approaches:
- With the best ones reported in the literature and
- With the ones currently used in Vale, trying to improve the quality of its outcomes
The final deliverable of this project are:
- Configurable data handling and modeling python scripts, developed in accordance with Vale´s architecture standards, and
- An accompanying technical report with documentation of the developed system as well as the findings and recommendations to be used in similar upcoming challenges Vale may face.
The candidate will have to:
- Perform a survey about the state-of-the-art in data science techniques applied to safety incidents prediction and convey findings via a written report
- Perform a comprehensive assessment of the applications within Vale’s AI Center that predicts safety incidents aiming to find improvement opportunities (about 5 models applied in about 10 different scenarios, possibly including text analytics)
- Develop phyton scripts to implement the proposed improvements while following the enterprise architecture standards and guidelines
- Run, test, fine tune, measure, and compare the prediction quality of the proposed scripts and procedures with the existing ones at Vale and convey or present findings via a comprehensive report
- Develop a minimum viable incident prediction product and apply it to test its effectiveness in a new area of Vale. Methodology/techniques to be used: Agile, CRISP-DM
The candidate must be a master's degree student in the field of data science or related areas with a bachelor's degree in computer science/data science/industrial engineering (Portuguese speaker- highly recommended skill).
The following hard skills are highly welcomed: Python programming language for data science, DevOps or MLops, previous experience in predictive models development which can be in a job or in an undergraduate capstone project.