LoadSmart

Probably the most important job of my career. The way I approach engineering today was shaped here.

The Gig

LoadSmart was a logistics tech company focused on freight and trucking. Someone needs a load hauled from A to B, they get a quote (via the public site or API integration for larger shippers like Coca-Cola), LoadSmart finds a carrier and handles the whole operation. The quote is final. That’s what the customer pays.

I joined as their first dedicated front-end engineer. Before me, front-end was a side concern handled by full-stack devs. My job was to change that.

The company was around 30 engineers when I joined. We all fit in one open office, which also meant some legendary Counter-Strike tournaments at happy hour. By the time I left, it had grown past 200, split across multiple rooms, gone fully remote during the pandemic, and scattered around the world.

The Product

Two sides: public-facing and internal.

Public site: thousands of clients per day getting freight quotes, requesting loads, API integrations for enterprise shippers. All of this went through front-end I worked on. 1,000+ loads handled every day.

Internal tool: the operations side. Sales reps and load managers working in it all day to keep those loads moving. What that looked like:

• Carrier sourcing: reps finding suitable carriers and contacting them through the app
• Load tracking: real-time map tracking once a carrier was assigned
• Finite state machine: I built the front-end for the FSM controlling every state a load moves through, from quoted to sourced to in-transit to delivered, with many states in between
• Task management: a full task system for reps tracking what needed to happen at each stage. A completed load required specific documents, a checklist enforced that. Hundreds of different task types total.
• Load reassignment: reps handing off loads to the right people for each part of the process

What I Fixed

The Monolith Escape

The frontend was tangled inside a massive Django monolith. Every deploy meant deploying everything. I led the migration to a separate modular React architecture with zero downtime. Hundreds of reps used the tool daily so we migrated piece by piece, live. Like changing the engine of a car while driving it down the highway.

Build Times

Build times were approaching an hour. I got that down to under 5 minutes through dependency optimization, build config tuning, and deleting things that shouldn’t have been there.

The SVG Vampire

SVG animations running off-screen, invisible but hammering CPU. Users’ laptops were getting hot. Tracked them down, killed them, fixed the underlying pattern so it couldn’t happen again.

Testing Culture

Code coverage when I joined: basically zero. When I left: over 90%, with tests that actually catch bugs rather than just inflate numbers. Setting up the culture and infrastructure mattered as much as writing the tests.

Performance and Reliability

Got all public-facing front-end to a perfect 100 on Lighthouse. Zero crashes across monitoring tools including Sentry and DataDog.

The Component Library

Built @loadsmart-data-visualization, a charting component library using Recharts, published to NPM. Still getting downloads to this day with no one maintaining it, which is both flattering and slightly terrifying.

The Experimental Squad

Toward the end, I was hand-picked into a smaller squad tasked with building solutions for a specific client segment: companies too big for the manual public site quoting flow but not yet at the enterprise API integration tier. A different kind of problem, smaller team, more autonomy.

Why I Left

I had done what I came to do. The front-end was in a good place, the team had grown, and I was ready for something new. No drama, just the natural end of a chapter.

Related

• @loadsmart-data-visualization - The charting library
• React - Primary framework
• Django - The monolith we escaped from
• CI/CD - Build pipeline work
• TypeScript - Introduced during my time there