> big projects
> portfolio
> contributions to a notable open source software project
This might help, it is something you can talk about on the interview. I wouldn't spend a lot of time thinking about if it was notable or not. Just go to Github, look for popular projects, perhaps in a language you might know somewhat like C# ( https://github.com/trending?l=csharp&since=monthly ), browse the project issues page and pull requests and see if the maintainer(s) apply patches in a timely manner (since you're a junior, timely can extend to within a few months).
Find a project that seems interesting, and an issue you think you might be able to handle, and try to fix the issue. Sometimes it's a not-too-technical person who just had the application crash or something and they report the issue. So pick a project and issue that look promising and try to fix it. If you can fix it within a week (or a day, or whatever), perhaps move on to another issue or project. Perhaps do this a few times.
At some point a maintainer will look at your fix. They might just apply it directly, or massage it into the project, or they may reject it. They may have some comments you might learn from.
I wouldn't spend a lot of time dwelling on it and thinking about which project is "notable" enough to contribute to. Better just to dive right in. Just get your feet wet and fix someone's problem. After that, then if you want you can think about which project is notable enough to work with and what looks good in a portfolio.
An internship would be essential if you weren't work as a "part time...junior fullstack web development(.NET stack and SQL)". Since you're doing that it becomes less important, and is not necessary and essential. An internship would not hurt and might help, but since you're doing real work, an internship becomes a little less important.
> I dont have any big projects done quite yet
Big projects start as little projects. For myself, I took a college course in Java for a few weeks in mid-2009, and then went off and did like I mentioned above - I chose a random open source project and fixed a bug, which was accepted. Then in late 2010 I spent two months writing a self-contained Java app which did something I wanted to do, I released it as open source. Along the way I sent a bug report to a similar Javascript app - I noticed two bugs while comparing my app to their app - I did such heavy QA on both apps I probably QA'd their app more than they did. Not long after that I started programming simple Android apps, and then more and more complex Android apps - this was in 2011. Last month my Android apps made over $3000 in revenue.
So I would say don't worry about "big projects" and "portfolio" and "notable" so much. Just start small and keep at it. Things can build on each other as well. I learned programming. Then Java. Then enough Android to do a simple app. Then I ported a Java library and put an Android UI on it. Then I ported a C++ library and put an Android UI on it. Then I ported an OpenGL C++ app to an Android OpenGLES app. Then I did that a few times and am still doing it. It would be hard for a .NET developer to port OpenGL C++ apps to Android. But I started simple, and then using what I learned as a base I did something else simple, then using that higher base I did something else simple, and now I'm doing something which would be considered slightly complex.
I've become interested in AI and machine learning and am starting a project in a simple way. Right now I am putting the data together and playing with scikit-learn. Once the data is in good enough shape I will start playing around with it and get more familiar with ML prediction. My initial model is pretty simple - I go to a few local supermarkets and buy out all their juice (5-6 bottles). Sometimes I go back and see if the juice is back in stock yet - sometimes it is, sometimes it is not. So I am keeping tabs on this and will see if I can get ML to predict the probability the juice is back in stock. It's nothing earth shattering, but it will save me a trip to the supermarket, or at least tell me which local one to go to. I have more ambitious ideas for other uses of ML prediction, but this is simple and useful and is a good first step for me. Who knows where it may lead?