My current activity is connected with the GPU/CUDA distributed processing of graphs on Ubuntu Linux. The job involves extensive use of C++, MPI, CMake, Git command line, Algorithms&Data Structures (especially Graph algorithms), and obviously GPU/CUDA. As part of the job, I read research papers on the relevant technologies.

Recently I implemented a raytracing application on the GPU: OptiX, OWL (OptiX Wrapper Library), RTX. I learned raytracing very quickly because I am familiar with GPU computations in general (CUDA).

I spent some time implementing vectorization (SSE2/AVX2/FMA) for ALGLIB: https://www.alglib.net/

One year was spent on developing a high-performance massively multithreaded networking application for web scraping. The application consists of multiple servers and even more microservices that communicate with each other via HTTP/REST and Windows named pipes. The applications are written in C++ (the main part) and Python (e.g. Django web interface, integration with Selenium and Tensorflow, etc.). Among the other technologies used there are libcurl for sending requests and receiving responses, oatpp web service framework for C++, nlohmann JSON library, dbghelp.dll for call stack, and minidump collection, MariaDB connectors for C++ and Python, etc.

1-hour demo: http://www.metacafe.com/watch/11919805/demo-of-software-comprehension-tools/
Website: https://www.softwarecomprehension.com/

We are using high-performance computing (CUDA, SIMD/AVX2, multithreading, NUMA/multiprocessing, algorithms & data structures) for large-scale calculations in the Insurance industry.