Work Stories

I Learned to Program When I was 7 years old

I wrote my first lines of Apple BASIC in 2nd grade. It didn't take long for me to become known as "the computer guy."

Those early years set the course for my entire career. Programming wasn't just a skill I picked up; it was the foundation of who I am as a builder and leader in technology.

Co-Op at PTI

Learning Protocols while Earning Money for School

As a co-op student, I wrote PTI's 2nd-generation DLL that carried all the SCSI protocol test code that I later expanded into my thesis. I also contributed to SCSI Toolbox, the company's flagship $5,000 per-seat utility. That's where I really learned the SCSI protocol from industry experts.

But PTI made a fatal mistake: they took a $1M custom development deal from one major customer and neglected their main product. Their only competitor took the lead. By the time PTI reacted, it was too late. The company split, and the half chasing growth collapsed.

I carried that lesson with me: don't neglect the core product to chase the side-quests.

Moving to Silicon Valley

Data Transit: Optimizing for 1000x Performance

On the recommendation of a PTI colleague, I joined Data Transit. It was my entry point into Silicon Valley and a turning point in how I thought about software performance.

I worked on PacketMaker SVT (Spec Verification Test), a tool used by storage manufacturers that sold for $30,000 per license. At first, it was painfully slow: a single command took 10 seconds to execute.

I partnered with another developer in a profile-fix-repeat cycle. I'd isolate the biggest bottleneck, he'd address it, and I'd profile again. We iterated like this until the program ran 1000x faster, from 10 seconds down to 10 milliseconds. That was my first case of systematic, team-based optimization.

I also represented the company at SAS plugfests at UNH, working directly with storage vendors to validate their hardware. And eating plenty of lobster at Warren's in New Hampshire.

Data Transit was eventually acquired by Finisar, but for me, the experience was a masterclass in both technical depth and industry collaboration.

Lessons in Process and Alignment

AMCC/3ware: The Power of Automated Tools

At AMCC, I built test tools for the 3ware RAID controller firmware team and designed a RAID configuration app in Objective-C. We even took it to Apple for design review.

The engineering side of the project went well, but we made a critical mistake: we thought we needed to present a well-polished app that people would love so we didn't include other stakeholders until late in the project. When we finally showed them the polished app, they felt left out and effectively sabotaged it.

That experience stuck with me: involve stakeholders early and make them part of the project.

I also had a front-row seat to history: AMCC sent me to Apple's 2007 Developers' Conference when Steve Jobs announced the first iPhone. I came home and told my wife I'd just seen the coolest thing ever. She bought me one as soon as they went on sale. (No, I didn't buy Apple stock and I'm still kicking myself.)

Soon after, my friend from PTI and Data Transit asked me to co-found SerialTek. I couldn't pass up that opportunity.

Co-Founding a Startup

SerialTek: Building an Industry-Leading Protocol Analyzer

As software manager at SerialTek, I had the chance to build something from the ground up.

In just 13 months, we designed and implemented an industry-leading SAS & SATA analyzer. I managed two very experienced engineers while writing much of the Java-based UI myself.

One of my proudest contributions was building automated testing into the analyzer. It captured traces overnight and verified traffic for correctness. My engineers loved it and automated testing was delivering value every day.

But startup life wasn't all wins. The product was so strong that our competitors' best move was to sue us. After three years with lawsuits, no raises, deferred salaries unpaid, and no written equity agreements, I walked away.

My two engineers left a month later. I passed on an opportunity to join them at WhatsApp. Instead, I chose to join old friends at Pliant, which later sold to SanDisk.

Things might have ended up different if we had been dreaming about the long-term vision rather than focusing on the temporary problems. A little hope goes a long way. Luckily SerialTek made it through the hard times and is still going strong today.

Saving a $1M Deal

Pliant Technologies: SSDs and Automated Testing

At Pliant, I took over and expanded automated testing using QualiSystems. I even earned my first LinkedIn recommendation for quickly mastering their tools.

I went further by centralizing SSD manufacturing and QA test data into an ASP.Net web dashboard. It gave engineers instant access to part-level details across every drive we produced.

That work paid off in a crisis. HP, our biggest prospect, ordered 500 SSDs for an evaluation. Some failed in early testing. If the design or firmware was flawed, the $1M deal was at risk.

I used the database to correlate failures against manufacturing data. A clustering algorithm revealed the root cause: 11 of 12 failed drives used SAS connectors from the same bad lot. I checked the findings with our statistician and it was confirmed: it wasn't our design; it was a bad batch of parts.

That finding saved the deal. When SanDisk acquired us, I was recognized as one of the "must-keep" employees.

Roller Coaster Ride

SanDisk: Scaling Teams and Elegant Solutions

When SanDisk acquired Pliant, I stayed and grew the tools team into a 12-person group. Our charter was automated testing and test data infrastructure.

I built a results dashboard that became indispensable, with 400 internal users collaborating on test and manufacturing data for every SSD we made.

When Operations needed a manufacturing database, I partnered with Microsoft to stand up a globally replicated, highly available system spanning three countries. I learned from top engineers at the Microsoft Technology Center where we stress-tested the system to handle 500+ simultaneous test machines.

We were scaling the enterprise SSD business to $B in just 2 years. Pliant grew from 40 people to 150 in a year and didn't stop there. That kind of explosive growth is like a thrill ride: it changes fast and has ups and downs, you learn fast, and it's a lot of fun.

Speaking of fun, I also got to travel to China, Japan, and Malaysia, touring the chip fab and manufacturing plants. I came home with a dozen ideas for improving our tools and tests.

Eventually, organizational shifts pushed me under a director I didn't mesh with. I was part of a reduction in force, but the timing aligned with finishing my Masters' degree and I was ready for the next chapter.

The Road to Autonomy

Georgia Tech: Computational Perception and Robotics

While working full-time, I completed Georgia Tech's OMSCS program, focusing on robotics under Dr. Sebastian Thrun and computer vision.

These courses expanded my computer science skills, adding computer vision, advanced algorithms, high-performance architecture, machine learning, and robotics to the list.

The experience cemented my passion for autonomous vehicles. I can't wait for the day when self-driving technology is ubiquitous for both the safety and the freedom it brings.

Creative Problem Solving

Lumenir: Building High-Availability Systems in Python

Lumenir was a short but pivotal stop while relocating back to Colorado. The company pitched itself as working on power delivery for the developing world, but much of the work was engineering consulting.

I got to build an embedded management system for YellowBrick, a SF Bay Area startup. I created it from scratch in Python, managing up to 16 blades with redundant server managers. It was highly available, robust, and proved the power of rapid development with the right tools.

The experience gave me a solid foundation in Python-based system design and earned me a LinkedIn recommendation from the customer.

Building Teams and Shaping Culture

PTx Trimble: Autopilot Steering Systems for Precision Ag

At Trimble, I finally found the intersection of my robotics training and real-world impact: Autopilot steering systems for Precision Ag.

I rebuilt the software team after some major org changes, hiring 10 strong engineers over two years. I created videos and instructions that reduced the onboarding time from 6+ months to 3.

My team reduced the backlog from 400+ issues to less than half of that. We added measurable value for farmers and earned a reputation for being a high-performing team.

Key contributions:

• Delivered 27 quarterly firmware releases (and counting): Each release was named after a different rock band in alphabetical order. I've delivered the whole alphabet and then some. That takes discipline, persistence, and dedication.
• Led critical OEM projects: I cleaned up messy processes for our largest OEM customer (CNH), turning around the relationship and earning a LinkedIn recommendation from the client.
• Hands-on problem solving: Traveled to customer fields fixing steering performance in Arizona, debugging harvesting on steep slopes in New York, solving row sensor issues in Georgia, and working with CLAAS in Germany.
• Built tools and automation: From Insanity (automated HIL tests to validate pull requests) to CANulator (a CAN bus simulator), I built tools that removed development roadblocks and raised engineering quality, leading a cultural shift around testing and risk management.