Here are some of the projects I built.

fuzzomatic

python

Many developers do not do fuzzing because it may seem difficult or is sometimes forgotten. However, it is a very effective testing technique that finds bugs at runtime.

Fuzzomatic lowers the barrier to entry and can automatically generate libFuzzer fuzz targets for projects written in Rust, completely from scratch. It uses OpenAI's API to aid in generating fuzz targets that successfully compile. Fuzzomatic found 14 bugs in the top 50 most starred parser library Rust projects on Github for just under $3 of OpenAI API costs. Read the blog post linked in the Github repository's README for more details.

Github: https://github.com/kudelskisecurity/fuzzomatic

oramfs

rust

Data encryption prevents others from seeing what the data is. However, when the encrypted data is stored remotely on a cloud provider, such as Google Drive or Dropbox, the cloud provider can still see how the data is accessed. For example, whether it is read or written to, and which parts of the data are accessed. In some cases, this can leak some information. ORAMs "scramble" or hide these read and write access patterns.

OramFS is an implementation of an ORAM file system written in Rust using FUSE. The ORAM scheme that was implemented is PathORAM. It is storage agnostic and will work with any remote host provider such as Google Drive, Dropbox, etc. that can be mounted as a local file system. Thanks to OramFS, not only the data is encrypted, but also the access patterns are masked to the cloud provider. This added level of privacy comes at a performance cost.

Github: https://github.com/kudelskisecurity/oramfs

vcsi

python

Generate video previews as an image. The generated image is a collage of multiple screenshots taken at various points during the video. These collages are called video contact sheets and are useful to quickly get an idea of what a video looks like.

Github: https://github.com/amietn/vcsi

For more open source projects, see my Github profile.