RedPine Signals (Ceremorphic AI)

Robotics Software Intern (Aug 2020- Feb 2021)

Red Pine Signals now acquired by Silicon labs California, is a semiconductor company that started operation in 2001. As a part of their robotics team I worked on NVIDIA's Jetson TX2 board. One of the major roles was setting up company codes and modifying them accordingly to work with arm64 type architecture of the TX2.
Also Created a method of bundling software packages for one of the robot applications and set up the robot to run headless.

RedPine Signals (Ceremorphic AI): Projects

We started off with the set up of the NVIDIA Jetson TX2 environment; involving flashing the board with NVIDIA’s SDK manager and installing the necessary environments.
The next step was to install various python libraries onto the board, to be used to create and execute a wide variety of programs. Libraries like NumPy, OpenCV, Tensorflow, and Keras, Pytorch were installed. A total of 5 python programs were made to integrate and check if the board can run them. These involved the aforementioned libraries.
Two methods of making executables were introduced, Pyinstaller and Makeself. Pyinstaller is to be used when dealing with smaller libraries since it packs even the python libraries. Whereas the Makeself method should be used when making executables with libraries that are much larger in sizes like Tensorflow and Pytorch since it does not pack the python libraries only the company libraries and supporting files. In the end, a technique to remotely access the Jetson Tx2 with the WLan was shown. How the board can be run headless from any laptop or computer connected to the same wifi.

RedPine Signals (Ceremorphic AI): Text

RedPine Signals (Ceremorphic AI): Pro Gallery