An app with a web UI that allows you to pass in a model and run the membership inference and gradient inversion attacks on the model, for the context of model privacy in federated machine learning
Developed a flexible command line tool which allowed for the detection (and removal) of high level features such as pockets, chamfers and fillets using Spatial's SDKs, and then allowed for extraction of data about those features into JSON, or visualised the detected features in 3D alongside the original model
A web based radio leveraging io_uring for the event loop, WolfSSL for TLS, WebAssembly for decoding Opus audio packets and the WebAudio API to play the audio and visualise it
Initially programming the devices at this company took multiple error prone manual steps, so I put together a few small circuits and some Python3 software to automate all of the steps, such that the building of the firmware to actually running the device took about 10 minutes vs an hour if not more
This was the operating system fundamentals coursework, in which I worked with a group of 3 other people to implement user space application support, scheduling and virtual memory management.
In this project I, with 3 other course mates, implemented a compiler from WACC to assembly and for the extension I implemented part of a reference counting garbage collector.
Generated a decision tree using the ID3 algorithm to determine the originating room of a signal in an apartment. Additionally, I trained a neural network on a dataset comprising house prices, enabling it to predict the median value of a house based on various attributes like the area's population and its distance from the ocean.
From a model described in a paper, I implemented a queuing model using OMNeT++, to simulate traffic through a generic network.