some things about me

I like to build tools that let people unlock their data and I like to teach people how employ maritime tactical helicopters. My current employment at Spear AI is an amazing overlap of these two worlds.

I don’t have a traditional education for computer science other than some basic introduction to programming for engineers using Matlab. Since then I’ve picked up various languages and skills out of necessity when building things.

When working on my Masters degree in 2011/2012 at the MIT Space Systems Lab I had to learn C to write a communication port driver on the SPHERES control system testbed. Then I learned enough C++ to add a module to the high fidelity simulation environment for that same test bed. Also while at MIT I had the pleasure of taking 18.085 Computational Science and Engineering from the legendary Professor Gilbert Strang which laid an amazing foundation and intuitive understanding of linear algebra that has proved extremely useful in this age of deep learning.

This same project was my first introduction to version control software. By 2011 git had been around a few years but the lab was still using subversion (svn). It wasn’t until years later that I was first exposed to git.

During my time in the Navy, I picked up both Javascript and Python for building tools to solve problems I came across. This was largely because on the locked down IT systems I could still run client-side Javascript, and the ground support laptops that my aircraft used came with a Python interpreter installed as part of ArcGIS!

One of those projects led me to deploying a static site onto Platform One. Using Platform One led me to learn about GitLab CI/CD pipelines.

I started reading Medium and subscribed to the Towards Data Science channel in order to get a slightly more comprehensive. That was great as a beginner to see the shortform articles describing various developer workflows and popular python libraries.

As I got closer to starting my second career I had seen references to dreaded computer science “data structures and algorithms” courses. I wanted to see what that was all about so I utilized free online educational resources like MIT OpenCourseWare to watch the recorded lectures of 6.006 Introduction to Algorithms. I find it amazing and democratizing that we have free access to world class university level classes like this!

While I didn’t study computer science in any formal school, I think a continual learning mindset coupled with trying to solve interesting problems can get you as far as you want to go.

some things I’ve built

BadMax (deployed at https://badmax.dso.mil/)

A tool for building Air Plan documents from ashore.

When I realized that my peers on the Carrier Airwing Staff were building Air Plan documents by drawing lines in PowerPoint to recreate the format generated by the ADMACS program onboard the aircraft carrier I knew I needed to do something about it.

Cronus

A lightweight tool to cleanly archive post-mission recorded data to S3 storage (AWS Snowball Edge) or local hard drive location.

Most tactical units that I’ve seen, both in service and since transitioning to a contractor role, are really bad at capturing their mission data and using it to build a bigger picture of their environment and their performance over time. Building a smooth workflow (that operators will actually use) for just simply saving any data generated by their tactical systems is a simple but necessary step.

PmaParser

A command line tool for batch processing Post Mission Analysis files from the MH-60R.

The platform in my aviation community would produce all sorts of recorded information. Much of it was only accessible via click-ops graphical user interfaces. This tooling was unable to support queries against a batch of missions. As a first step in unlocking the Post Mission Analysis data file, this parser dumps the contents into standardized formats (e.g. CSV) instead of the native custom binary packed data format.

WaveBreaker

A tool for planning JDAM impact points to cover the area of uncertainty from a moving, unobserved target.

During my participation in the USAF Bravo Hackathon 01, while the focus was on building machine learning systems with some classified data sources, I realized there was an interesting tactical challenge that could be addressed by a good application of some plain old maths plus a good user interface. I decided to break with the pack and build this tool that could be loaded onto aircrew tablets as is and be used to make an impact.

some things I’ve written

some places I studied

  • S.M. in Aeronautics and Astronautics, MIT, 2013
  • B.S. in Aerospace Engineering, USNA, 2011
  • Machine Learning Certificate, Stanford Online, 2021

some things I’m okay at

I dabble mostly with Python and Rust, and mostly in the context of data processing, parsing, and storage.

I’m familiar with the Postgres database ecosystem.

Every now and again I’ll work into the Next.js ecosystem for some integration, but I know that’s not really my swimlane.

I’m intimately comfortable on the linux terminal. I’ve spent many an hour ssh’d into a remote server building and deploying code.