- Accelerating existing algorithms using CUDA and GPU.
- Minimizing hardware resource usage, e.g. GPU memory footprint, through proper resource management.
- Building GPU-accelerated Machine Learning Library.
- Test, bench, and build tools for benchmarking ML algorithms on the hardware accelerator.
- Design and develop new GPU operators with high-quality codes.
- Analyze the software and identify potential performance bottlenecks.
- Enhance current applications to better utilize hardware accelerators.
- Collaborate with other teams of engineers to provide accelerator related support.
- Conduct research on various neural network acceleration.
Experience & Skills Required:
- MS/PhD in Computer Science or other related fields
- Strong programming skills in C++/C
- Solid understanding on Computer System, Computer Architecture
- Hands-on experience in at least one of the following fields:
- DL workload programming & optimization on various hardware accelerators, such as GPU and FPGA
- Developing machine learning library
- Entry-level machine learning knowledge
- Excellent communication and interpersonal skills
Preferred Skills and Experience:
- Development experience in following open source Machine Learning Libraries (MXNet, TVM, Caffe)
- Experience in arm-based embedded systems/Nvidia Drive products
- Strong System Skills (Operating System, Parallel Computing, Distributed Computing)
TuSimple is an Equal Opportunity Employer. This company does not discriminate in employment and personnel practices on the basis of race, sex, age, handicap, religion, national origin or any other basis prohibited by applicable law. Hiring, transferring and promotion practices are performed without regard to the above listed items.