Research Engineer (Differentiable Simulator)

Job Description

The research project involves the development of high-performance computing and differentiable simulators for robotics learning. The research engineer is expected to contribute in the following aspects:
a)  Develop differentiable physics-based simulators for robotics learning and manipulation, including rigid body dynamics, soft coupling and fluid simulation.
b)  Build high performance computation forward simulation and differentiable simulation models on GPU, and optimize performance on forward and differentiable methods.
c)  Design control policy on robotics learning combining differentiable simulator and other learning methods, exploring the connection between differentiable simulation and large-scale robot deployment.
 

Qualifications

•    A bachelor degree in a relevant discipline, e.g. electrical / electronic engineering, computer engineering, computer science, etc.
•    Full experience in computer graphics and robotics area, especially in physics-based simulation, differentiable simulation, high performance computing.
•    Familiarity with auto differential libraries (JAX, AutoGrad, Warp, Taichi…)
•    Experienced on GPU programming, familiar with cuda thread scheduling, allocation, and optimization, have management and project planning capabilities for large-scale CUDA engineering and the CudaGraph.
•    Software programming in C++, Python and knowledge of Cloud Computing, Edge Computing.
•    Excellent analytical and computational skills.
•    Understanding of software engineering principles, preferred worked experience on open-source software.
•    The candidate must have the ability to work independently as well as part of a team with strong initiatives and have the curiosity to explore the unknown.
•    Open to Fixed Term Contract.

More Information

Location: Kent Ridge Campus

Organization: College of Design and Engineering

Department : Electrical and Computer Engineering

Employee Referral Eligible: No

Job requisition ID : 25891