Roadmap Data Science

[aacecandev]

The big problem is that Mairror is focused in two main areas:

• GANs
• Supervised Learning

The cool thing is that this project's name is not concise but open enough so it can be thought of as "you show something to your camera or send a picture (or a set of pictures), and then something happens."

• Redo the dataset cleaning
• Try different approaches for image face recognition (including real-life detection) (RCNN | YOLO | OpenCV)
• Try different neural network frameworks (Tensorflow2 | Pytorch - Pytorch Lightning | XDGBoost)
• We actually didn't implement GANs (i.e StyleGAN | DCGAN)
• Converting Jupyter code to Python code to make it automatically available to trigger a model training, code testing, code quality, etc.
• Data visualization in Jupyter
  • Concepts to learn to properly curate and know about the data that the dataset contains
  • Tools (matplotlib | seaborn | metabase | plotly)

CUDA

• General C++ knowledge
• General CUDA development knoledge
• Time sharing
• Tensor cores utilization
• GPU performance and utilisation