Advanced Methods for Time Series Data Augmentation- QC-384Project type: Research
Desired discipline(s): Engineering - computer / electrical, Engineering, Engineering - mechanical, Engineering - other, Computer science, Mathematical Sciences, Mathematics, Economics, Social Sciences & Humanities
Company: Lockheed Martin
Project Length: 6 months to 1 year
Preferred start date: As soon as possible.
Language requirement: English
Location(s): Montreal, QC, Canada; Canada; Canada
No. of positions: 1
Desired education level: CollegeUndergraduate/BachelorMaster'sPhDPostdoctoral fellowPreferred institutions: McGill University, Polytechnique Montréal, Université de Montréal, University of Alberta, University of Manitoba, University of Montreal, University of Toronto, University of Waterloo
About the company:
Lockheed Martin Canada is the nation’s most successful Combat System Integrator. We leverage modern engineering practices and technologies to design, integrate, test & deliver combat systems onboard some of the world’s most advanced warships. We are seeking highly driven and talented candidates to join our Emerging Technologies Research Group.
Describe the project.:
Our Research Project will investigate the most recent advances in deep learning (such as GANs, LSTMs and Autoencoders) with the objective of building a multi-purpose time series data augmentation engine.
In pursuit of this goal, the ETR Group will be tasked will the following objectives:
- Conducting a detailed literature review, researching the most widely-used methods for synthetic time series generation;
- Surveying available open source tools and packages;
- Amasing a collection of open source time series datasets from a variety of application domains upon which to conduct trials;
- Conducting experimental design, establishing a clearly defined methodology and measure of succes;
- Developing deep learning models (including design, training & optimization) and performing rigorous statistical testing against outputs;
- Producing documentation & delivering findings to senior management.
The ideal candidate will have a strong working knowledge of both traditional time series analysis and deep learning methods. Through their involvement in this Research Project, they will work to transform the field’s latest theoretical breakthroughs into practical engineering tools. Required skills & experience include:
- Proficiency in Python & past experience with deep learning frameworks (preferably TensorFlow and Keras);
- Proven research experience, demonstrated through education, previous employment, research or personal endeavors (candidates are encouraged to include links to published articles, preprints, blog posts and code repositories);
- Effective communication skills with both technical and non-technical members of the engineering and management staff;
- High degree of initiative, autonomy and drive.
NOTE: Canadian citizenship is required as any successful candidate must be eligible to obtain a Controlled Goods Program security assessment.