Bailey Johnson — Software Engineer, Applied AI

Hi 👋, I'm Bailey.

I'm a software engineer at Flexport working on agents and search. I have experience building services and APIs that power complex applications. I like handling software integrations and really dig distributed systems.

I've spent a large portion of my career working with machine learning models. First by building the infrastructure and tooling necessary to support them, and more recently by leveraging frontier models to build powerful client-facing features. Now we literally work with them every day.

The amount of progress these last few years has been astonishing and has made me incredibly excited for what comes next in the intelligence era. I'm currently studying ML at Georgia Tech to better comprehend how we got here and where we're going.

Quick Career Overview

In May of 2020 🦠, after a short internship at DNSFilter (2016 - 2017) and about a year and a half of co-op dev work at WetStone (2018 - 2020), I graduated from Coastal Carolina University with a B.S. in Computer Science. I then relocated from South Carolina to the Houston-area to work full-time as a SWE at NASA's Johnson Space Center 🚀.

A little over a year later I accepted a fully remote position with Press Ganey (NarrativeDx). I grew a lot in that role and it also afforded me the opportunity to work from anywhere in the states.

I visited Seattle in February of 2022 and fell in love with it, so eight months later I packed up my things and drove across the country. I joined Flexport in April of 2023.

More About Me

I've been a computer geek pretty much my entire life. In middle school I was building webpages, learning TI-BASIC and writing some pretty bad games in Python. I had the opportunity to join a CS program during high-school and also started my first internship senior year.

While I've landed on ML and software engineering, I was pretty interested in cybersecurity growing up. It led me to some security conferences, CTF-events and eventually to me landing that co-op position with WetStone (doing development to support their security/digital forensics tools). I think it was there I realized I was less interested in security professionally and really just liked building software.

When I'm not at my desk I enjoy live music, hiking, longboarding, playing guitar and making coffee ☕.

Things I've Built 🛠

I've had the opportunity to contribute-to and lead the development on some pretty neat projects throughout my career.

Ask Me About:

Building Flexport Intelligence
Taking a NLP Pipeline from a Monolith → SOA
Supporting ISS Operations Above FCR-1 in the Mission Control Center

For more information about these and other projects I've worked on professionally, please check out my resume and don't hesitate to get in touch.

Hobby Projects

The speedup provided by agentic development has made it much easier to iterate on ideas, both at work and at home. Most of my recent hobby projects have been byproducts of that, or half-baked things I was finally able to finish implementing without as much of a time commitment. Don't get quite as much time as I'd like for stuff like this with grad school but here are some of my favs.

darkreel 🍿

A self-hosted app for storing and streaming media with end-to-end encryption. An exercise in private, secure and performant agentic software development. Pairs well with PPVDA and darkreel-cli.

darkreel.io View Source

Go JS

spike-ensemble

A closed-loop system that drives biological neurons on a Cortical Labs CL1 chip to produce real-time generative music. Musical context (rhythm, harmony, dynamics) is encoded as electrical stimulation; spike output is decoded into notes. A coherence-based feedback loop shapes the neurons toward musicality over time.

spikeensemble View Source

Python CL SDK

havenwood

An animal sanctuary management game built with Phaser 3. Rescue animals, build habitats, hire staff, research upgrades, and attract visitors. All sprites and sounds are generated procedurally at boot.

havenwood.io View Source

JS Phaser

Old Stuff

These are some of my favorites from further back, built either for undergrad courses or for fun. All of these are available on my GitHub.

balance-table

This project implements a balancing algorithm for placing 'blocks' (integers) on/into a 'table' (two-dimensional matrix), ensuring the calculated center-of-gravity is not to exceed a range at which the table would fall. It's a Python application that makes use of a library called Eel for building web GUI's (the GUI allows users to input a table size and visualizes the placement algorithm in real-time).

View Source

Python Eel JS

the-wall

The Wall is a single-page app I built with the History API to demo client-side routing without the use of any framework or designated router. It's not connected to any data store, but does support a version of route-based searching for posts by ID.

View Source

JS History API Bootstrap

mpi-matrix-mult

This project uses C + MPI to perform matrix multiplication in parallel. The project folder contains scripts that generate the matrix files, which are then read in by serial/parallel multipliers. Built for an intro parallel/HPC course, performance benchmarked on XSEDE's COMET cluster.

View Source

C MPI