Why I’m Working on This

This is how I got here.

I didn’t come to nutrition through a conventional career path.

I came to it through family, sport, ethics, technology, and eventually science.

My father struggled with type 2 diabetes. Watching that made diet-related chronic disease personal long before I studied it academically. It also left me with a question that has stayed with me:

Why is knowing what would help so different from being able to sustain it?

At the same time, endurance sport and adventure had become a major part of my life. I ran, cycled, raced triathlons, and climbed mountains. Nutrition interested me initially for a very practical reason: performance made its consequences tangible. Fuel well, recover well, adapt well—or pay for it.

Food was becoming important to me for another reason too.

Concern about environmental sustainability and animal welfare led me first to vegetarianism and eventually to a completely plant-based diet. I wanted to know whether the way I ate could align with my values without compromising what I asked my body to do.

Over the years, that became more than a personal dietary choice. I summited Denali, completed multiple full-distance Ironmans, and at 49 ran a marathon in under three hours—along with ultradistance races, dozens of Colorado 14ers, and three Boston Marathons. I trained for and completed all of it while eating plant-based.

But there was an important limitation:

I had experience, not scientific authority.

My own performance could demonstrate that something was possible for me. It could not tell me why it worked, whether it was optimal, what the tradeoffs might be, or whether anyone else should eat the way I did.

I loved the adventure for its own sake, and I was consciously trying to lead by example. But the distinction between what I could demonstrate and what I could legitimately claim increasingly mattered to me.

So I went back to school.

I brought an unusual background with me. Before entering nutrition, I had spent years in software, telecommunications, entrepreneurship, business development, and operations. I had seen what technology could accomplish when it solved a genuine problem well, and I entered nutrition wondering whether those capabilities might eventually have a useful role in the field.

Later, on scholarship, I completed MIT Professional Education’s 80-hour No Code AI and Machine Learning: Building Data Science Solutions program.

What I still didn’t know was what the right problem was.

At LSU, nutrition became less about belief and more about evidence. Biochemistry, physiology, food science, nutrition assessment, counseling, and research methodology taught me to question assumptions—including my own.

The more I learned, the less interested I became in technology for its own sake.

The question was never where can AI be inserted into nutrition care?

It was:

Where is there a real problem worth solving—and can technology help without weakening the expertise, relationships, judgment, or autonomy already there?

That path included work supporting the Nutrition for Precision Health study at Pennington Biomedical Research Center. That experience helped shape my questions; it does not imply a current affiliation or endorsement of KeyRD.

The premise of precision nutrition is compelling: understand individual differences well enough to make nutrition recommendations more precise and effective.

But working in clinical research sharpened another question for me:

Even if we become much better at determining what someone should do, how does good care survive the weeks between clinical encounters?

A clinician can perform a careful assessment, develop an appropriate intervention, and explain it well.

Then the patient goes home.

Work happens. Family happens. Travel happens. Fatigue, cost, symptoms, conflicting information, changing schedules, motivation, uncertainty, and ordinary life happen.

The care plan has to survive all of it.

I explored that problem further through a systematic review I led examining mobile-health interventions designed to improve adherence to dietary prescriptions. We screened more than 7,000 records and ultimately included 39 studies. The manuscript has been submitted to JMIR mHealth and uHealth.

That review gave me a much clearer view of what this category has—and has not—demonstrated.

The literature showed promise, but also substantial uncertainty. Technology could help. Simply adding more notifications, engagement, or automated advice was not the answer.

That distinction mattered to me.

More intervention is not necessarily better care.

More engagement is not necessarily better adherence.

And a more capable AI system does not automatically acquire the authority to decide what should happen next.

Nutrition care is largely organized around encounters. Nutrition itself is lived continuously.

That is the problem behind KeyRD.

KeyRD is the current research and development project at Statera Systems, the company I founded on a simple premise: technology should extend human capability without unnecessarily taking control away from the people responsible for the work.

KeyRD applies that premise to a specific question:

Can technology help preserve the intent of nutrition care between visits without replacing the dietitian or creating another engagement machine?

The approach is deliberately restrained.

Sometimes the right action is support.

Sometimes it is a question.

Sometimes the clinician needs to be brought back in.

And sometimes the right action is to do nothing.

The goal is not maximum intervention. It is the minimum sufficient intervention required to keep good care connected to real life.

I designed the prototype to make those ideas inspectable. It includes plan interpretation, boundaries on proposed actions and records that let a reviewer examine a decision. The public Maya demonstration illustrates how those ideas might support a longitudinal study; it does not run one.

I’m seeking clinical and methodological collaborators to ask harder questions: can we define and measure adherence well, can we execute a bounded intervention reliably, and does adaptive support help more than a simpler alternative?

The separate continuity-evaluation plan asks whether proposed decisions are useful to an RDN. It cannot answer whether a JITAI changes dietary adherence. Those questions need different designs and evidence.

Designing KeyRD gave me something concrete to investigate. It did not establish the answer.

I hold a BS in Nutrition & Food Science from LSU and completed the Didactic Program in Dietetics. I am not an RDN. I founded Statera Systems Inc. and am responsible for KeyRD’s concept, research, and product and architecture design, defining its operating boundaries, and ensuring that the claims we make about it do not run ahead of the evidence.