Travis Lewis

I have always been drawn to problems where the answer has to be right, not merely persuasive. That started early, with mathematics at Choate and later at the University of Pennsylvania. A system either works, or it does not. Numbers do not negotiate.

After school I built my career in mission-critical technology, on trading desks where failure was expensive and time mattered. My job was to understand the operating requirement, turn it into architecture that could survive pressure, and keep it reliable when money was at risk. That taught me something I still use. The purpose of any system is to serve the outcome, and the person responsible for it has to trust what the system gives back.

I eventually left that world and founded Fulcrum Solutions, building computational infrastructure for cryptocurrency mining. Different domain, same discipline. Design a system that works under real conditions, monitor it, and improve it. From there I became a professional trader, and that is where I found the single weakest component in any strategy. Human emotion. Attachment to outcomes. The conviction that you are right when the data says otherwise. Discipline fails at the point where it matters most, and it fails because of the person, not the method.

So I wrote my trading philosophy into code and took myself out of the loop. The algorithm executed what I had designed, without the emotional interference. It made money.

Then Artificial Intelligence matured enough to analyze my algorithm, so I fed it in. It rewrote the strategy and improved it. A system I had spent years refining was made better by a machine in a fraction of the time. I could have been threatened by that. Instead I chose to understand the capability completely, and spent hundreds of hours studying how language models reason, where they fabricate, what they need to perform well, and what they need to stay responsible.

What surprised me most was that this was not new at all. Augmenting human judgment with technology is a century-old tradition, Intelligence Augmentation, where the human stays the hero. I had arrived at something that already had a name and a history. Most people building with AI have never heard of it.

Here is the reconciliation I had to make. In pure-pattern trading, removing the human wins. The emotion is noise, and the algorithm is better without it. But the decisions that matter most are not pure-pattern problems. They involve values, relationships, history, ambiguity, and judgment no machine can replace. There, the system's job is to remove the noise so your judgment can be at its best. It does not replace you. It prepares you.

Avant Garde grew from that understanding. I build governed intelligence systems for people whose responsibilities span decisions, commitments, relationships, and time. The tradition beneath the work is Intelligence Augmentation. The infrastructure may connect the major systems early, but the deeper value develops through use. The system earns trust, you get better at directing it, better practices emerge, and both judgment and capability grow stronger over time. The system prepares the record, preserves continuity, and strengthens the person. Judgment stays yours.

They sell you the tool. We install the system.

Travis Lewis

Founder and Chief Cognitive Architect · Avant Garde International Group