I’ve decided to get Coursera’s GPU Programming Certification. This will allow me to make very deep changes to deep learning models that go beyond stacking layers and activation functions and their connections together. This could also allow me to work on shaders if I ever get into game dev.

Deliverables

I don’t want to just complete this course, I want to apply it. So here’s what I want to accomplish:

1. Complete the Certification.
2. Create a deep learning model with pre-existing weights that uses one or more layers that uses my GPU code. This will probably be run on my desktop since I have an NVidia GPU on it.
   • Using pre-existing weights is important here. I want to see if I can improve the performance of a pre-existing model by making changes in the GPU.
   • Whether it actually is faster/more accurate/somehow better than the existing model doesn’t matter. It needs to run at least as well as the default model on my system.
3. Complete all of this by the end of February. Have multiple things I need to work on to get my next job so I can’t allow this to go on and on.

I considered adding something like “run GPU code on Android” but that would be going beyond the scope of the certification.

Learn More

The certification is divided into 4 courses.

• Introduction to Concurrent Programming with GPUs. Coursera / My Posts. I’ve finished this course, so here is My Posts / Coursera Class Page.
• Introduction to Parallel Programming with CUDA. Really looking forward to this course. My Posts / Coursera Class Page.
• CUDA at Scale for the Enterprise. Here is when I should be able to begin Deliverable 2. My Posts / Coursera Class Page.
• CUDA Advanced Libraries. Since this has a portion focused on Machine Learning this should be able to make Deliverable 2 easier to accomplish. My Posts / Coursera Class Page.

Here is the most recent post in this category!

• 
  GPU Programming 6: Non-Global MemoryFebruary 6, 2026
  The last two modules of Introduction to Parallel Programming with CUDA go over the other major types of memory and how to interact with them. I’m also going to add a bit of my own commentary and things that weren’t fully explored in the course. Memory Overview Let’s look at a bird’s eye view of… Read more: GPU Programming 6: Non-Global Memory