Brady Pomerleau

Here I acted as the primary designer and end-to-end developer of a system for creating dynamic networks of communication devices. We tested interconnective features for these devices and had to move devices to connect them with custom cabling.

I designed a network of managed switches that allowed us to create and destroy network topologies dynamically without moving physical hardware. I wrote a REST API using Python Flask and Apache to accept YAML topology files. The network topology specified in the file was then generated by creating VLAN tunnels on the managed switches. Available resources and active tunnels were tracked with MySQL databases.

Later, I developed a command-line interface for the system using Go. The CLI could create, list, and destroy active networks. This interface would cache canonical topologies and reference them by name instead of supplying the file each time you wanted to create.

This dynamic network generator allowed us to double the amount of automated testing time available overnight and was critical for the project's success through the COVID-19 pandemic.

A personally motivated project, a gimbal for mounting a DSLR camera for time-lapse and panorama photography.
I designed and prototyped the entire project, including mechanical, electrical, and software.

The gimbal was designed to allow rotation around the optical center of the camera instead of the center of gravity. It consumes more power but reduces parallax shifts.
I adapted an open source DSLR control library to control the camera via Arduino microcontroller over USB.
I built a hot-swappable power converter to power the gimbal with power-tool batteries.

I developed a reverse-proxy server using Apache2 on RHEL.
The reverse proxy was implemented to provide software services across multiple segments of an enterprise network.
The server included a front-end landing page implemented in Python Flask and Bootstrap. The page had links to various servers proxied by my reverse proxy.
Authentication to access the server used OAuth2.0.

Using the API provided by the IBM CLM suite, I wrote several Python applications to retrieve, analyze, and inject data into the CLM software.
Some examples of collected data include data from network analyzers, test logs, and unit and integration test results.
Data retrieved from the CLM suite was used in automatic report generation.