I've applied to grad school multiple years running, and haven't found success, probably because software infrastructure isn't considered publishable research (although I've also published a novel distance metric for protein concentrations in cytometry analyses). I'm currently living off my savings and using the free time to follow through on long-term projects like:

1. Making pip the fastest resolver in the west. This began with pypa/pip#7819 and I got enough people excited to release pip install --report and extreme caching improvements, summarized in my upcoming packagingcon talk.
2. Exposing a build.rs interface for spack. This makes it much more reliable to include high-quality C/C++ codebases in Rust code, without pkg-config. A tentative prototype which integrates the emscripten compiler to include that C/C++ code in WASM builds is planned next: see ffmpeg-rs.
3. Releasing a Rust library and CLI extending the Signal protocol for general message encryption to replace gpg. Signal themselves doesn't have the bandwidth to expose their crypto for more general use cases, but I do: signalapp/libsignal#467. The project is called grouplink and I will be releasing a version 0.1 before 2024.

If any of the above would improve your programming experience or help with your actual job, please consider exchanging money for goods and services. I check DMs on twitter (@hipsterelectron) and mastodon (@hipsterelectron@circumstances.run); feel free to ask for my protonmail or Signal. I TAed students for 3 years of undergrad, and I write better docs and test cases than anyone else I've ever met.