Throwing stuff at the wall
Right now as I’m writing this. I’ve been trying to use my programming and data science skills. To get some type of income. But as I don’t know how I would use those skills to get an income. I have been trying lots of new projects to see if they have potential. The most recent project was me implementing a research paper which made labels from seismic images. Which can be used for carbon sequestration. Granted I did not succeed but that’s for another article.
But hopefully, as I try all these projects. My skill level should improve. And I could be making more interesting projects. And if one of my projects becomes popular I can make a business from selling the software. But that’s mainly a long term goal. So right now I’m in the building stage of my expertise. Also trying to find ideas that can prove useful to somebody. Which was why I ended up on the seismic imaging. As I can imagine an energy company can use the software to find places to spot carbon storage areas. But when the project finishes I need to spend time again looking for another idea/project to work on. Which is something difficult as I try to find projects that are interesting and develop my skills. Basically not bog-standard tutorials like MNIST or titanic dataset.
But as deep learning and software engineering is wide a field. Any industry you can think of can use some type of ML model or software. The only threshold that it must be somehow useful. To find a new project idea I tend to do a lot of googling. Of topics, I’m thinking about at the moment. Or project ideas I have written about earlier. After I come up with an idea I write it down. Then I start researching how to do it. For the seismic images, the idea was saved earlier. So I just had to google it. I found multiple research papers. And skimmed through to see the one can I can do. From there pick resources that I could use for the project. For the seismic image project, I read the papers more thoroughly this time. To get an idea on how to build the project.
For other projects, I may have read other peoples blog post talking about how they build a similar product. Or an element of my project. Then I could start building a small prototype from there. After I build the prototype, I can optimise for things like the design. Or how the code is written. Depending on the project. I will want to ship the project. So I may use something like a docker-file which I used for the carbon dashboard project. Then publish on a cloud provider. Right now I’m using google cloud. But I have been thinking to use Microsoft Azure. After I do a write up of the process of how I made the project. And what I’ve learnt and how I can improve. As normally is something incomplete in these projects. As they are supposed to be a prototype.