Postdoctoral Research Associate, High-Performance Parallel Graph-Based Machine Learning: University of Waterloo, Canada

Εικόνα stratis
Κατηγορία: 

 

Description

We are looking for a postdoctoral research associate to join our research group at the Computer Science department at the University of Waterloo. Our goal is to develop parallel and communication efficient algorithms for large-scale graph-based machine learning.

Example algorithms and applications include but not limited to: optimization algorithms for local graph clustering and optimization algorithms for training graph neural networks for node classification, link-prediction, community detection, graph classification and graph generation.

The successful candidate will also collaborate with the Waterloo-Huawei Joint Innovation Lab and with the end goal to develop a new system or support existing systems on graph-based machine learning. The candidate will also be part of the Scientific Computation Group and the Waterloo Artificial Intelligence Institute.

Starting date: January 1, 2021 or later.

Duration: 1 - 3 years (ignore the 1-2 years "Job Duration" to the right)

Salary is in CAD.

Additional information can be obtained by contacting Dr. Kimon Fountoulakis by email at .

Apply through SIAM or by sending an email to Dr. Kimon Fountoulakis following the instructions in here.


 

Requirements

PhD, within the last 3 years, in one of the following or other relevant subjects: numerical optimization, scientific computing, parallel computing, applied math, machine learning.

Knowledge of (randomized) first-order numerical optimization algorithms, numerical linear algebra, neural networks or other machine learning models. 

Experience in code development and computational experience in using high-performance parallel computing resources. In particular, demonstrated coding experience in C and/or C++, experience in parallel programming using GPUs and/or using tools like MPI and OpenMP. Experience with neural network frameworks such as PyTorch.

Publication record in high-impact journals, top-tier machine learning, and related conferences.

Excellent written and oral communication skills.

 

Startup Growth Lite is a free theme, contributed to the Drupal Community by More than Themes.