Customer Engineer
CurrentMy primary role includes handling customer-reported issues. As part of the process, I have to communicate with the customer to understand the case, collect logs and other data for internal reproduction. Test environment ranges from single multi-GPU system to multi-node clusters with Infiniband interconnect. Running different flavors of Linux executing applications involves machine learning framework (PyTorch) and HPC applications (GROMACS, LAMMPS, NAMD, and OpenMM). Here I have an excellent opportunity to learn different ML/DL algorithms from simple CNN/RNN to LSTM/BERT. Also, I spent time on the backend, running kernels on GPU and optimizing them. I had exposer to MPI/UCX/NCCL and other communication protocols. I was able to build custom docker and singularity images with needed packages.