Optimize the inference performance of various DL models (CNN, NLP, etc., )

• Identify bottlenecks in performance and propose scalable solutions

• Benchmark & Low-level profiling to analyse hotspots in order to exploit available bandwidth & computational power across cores

• Leverage hardware resources & distribute workload using parallel computing to get optimal performance

Qualification:

• Bachelor of Engineering/Bachelor of Technology, Master of Engineering/Master of Technology.

• Must have good programming skills in Python

• Very sharp analytical & problem-solving skills

• Exposure to Multithreading & parallel processing concepts

• Knowledge on Parallel Computing - CPU/GPU & Parallel Programming - OpenMP, CUDA

• Good knowledge of computer architecture & micro architectures like x86.

• Nice to have knowledge of linux/windows profiling tools

• Nice to have DL framework knowledge (Tensorflow, Pytorch, ONNX) & exposure to DL model architectures