TTU HackWesTex 2026 Developer Challenge: macOS AI File System Metrics
One the hidden challenges of running AI models on non-cloud infrastructure is the ability to support
the scalable data storage requirements of the users. Healthcare professionals, law firms and
other private companies may store terabytes or petabytes of data at rest in structured file systems.
Managing these file systems is largely relegated to mixtures of basic open source tools and highly
proprietary/file system specific tools designed to run in Linux or BSD (see ZFS pooling).
Recently, developers and users of private AI infrastructure have discovered the power efficiency and performance of Apple Silicon to run local AI models. Workflows such as OpenClaw and Exo have become commonplace is small organizations seeking to free themselves for cost prohibitive cloud-based AI tools. However, there are minimal tools available to manage and monitor file systems on macOS systems.
This challenge is designed to build a set of management and monitoring tools natively developed for macOS to manage file systems. We challenge the students to explore ideas and implementations for management utilizes that:
- monitor the health of file systems, including the backing block storage health
- monitor the performance (GB/s) of the file system I/O performance
- monitor the capacity and per-user quotas present for a given file system
- display the metrics in a useful form the administrators
- provide reporting mechanisms to alert administrators of nefarious users and/or capacity concerns
- support local (APFS) and shared (NFS,pNFS) volumes
- any other interesting file system related metrics, data, etc (the more the better!)
Challenge teams can use mixtures of any programming models natively available on macOS. This includes
Python, Swift, C++, C, Rust, Go, etc. All source code and documentation must be available open source
at the time of submission. The source MUST include instructions to build and execute the tools.
Makefiles and/or CMake scripts are highly encouraged in order to ensure portability. Apple XCode or
MS VSCode is not required, but can be used to assist in the development.
The final deliverables should be:
- The repository of source code, including basic documentation to build + execute the project
- A short demo of the functionality
- Additional points will be given for projects that also outline "future work" to expand the scope of the tools
Judging will be done Sunday September 13th at the end of the event. The winning team will be presented with a prize (to be announced at the start of the event).
The source code for the project must be stored in a publicly available repository such as Gitlab, Github or Bitbucket for judging. Source code storage locally on the teams' personal devices will not be accepted.