Story

I grew up as an only child in a refugee household in New York City, trying to understand how the world works. I saw how often institutions reward signals before substance. That shaped my interest in education, credentials, and human potential.

At Bronx Science, academic expectations were high. Policy debate gave me a method for thinking under pressure: examining claims, evidence, incentives, tradeoffs, and competing interpretations of truth.

After undergrad, I learned and then taught foreign exchange trading. I began building early artificial neural-network models for forex pricing, an introduction to the gap between a model's predictions and the decisions people make from them.

I helped build electronic financial exchange infrastructure at Panafex, the Pan African Financial Exchange in Uganda's Lake Victoria Free Trade Zone, and later at Caufex in the Republic of Georgia.

After earning my Juris Doctor at Benjamin N. Cardozo School of Law, my work expanded across hedge fund analysis, family office operations, global supply chains, and international energy-policy projects.

Those experiences taught me to pay attention to the rules and feedback that govern a system: what gets measured, who is accountable, and how decisions change when better evidence arrives.

At Koherent, a family office and private equity holding company, I served as Chief of Staff, Chief Operating Officer, and Chief Investment Officer. I worked with the chairman across investment, technology, and operations, turning priorities into plans, management processes, and decisions.

The work ranged from deal intake, diligence, and portfolio reporting to vendor coordination and cross-border execution between New York, Seoul, and international partners. I prepared executive materials, established recurring reviews, and resolved issues across functions.

That experience shaped my approach to operations: understand the numbers, make responsibility clear, and stay close enough to the work to know when a plan needs to change.

In markets and operations, performance is measured continuously. In education, attendance, assignments, grades, and credentials often stand in for evidence that someone has learned. I started asking what a learning system would look like if it took measurement and feedback as seriously.

I founded Virtu Prep to build that system. It is designed around learning science, AI, active practice, adaptive sequencing, and evidence intended to make learner progress inspectable. Its intended settings include schools, academies, test preparation, workforce programs, and other specialized learning environments.

As founder and CEO, I own planning, budgets, hiring, vendor strategy, and investor communications.

Through the company's first five years I built and led a product organization of 19: 11 engineers, 3 designers, and 5 QA. As the organization changed shape, I took on more of the technical work directly and rebuilt delivery capacity around AI engineering systems. That shift is why the engineering records on this site are personally attributed.

I coordinate engineering, content, design, and partners, while remaining hands-on in the software and AI workflows behind the platform.

In Virtu Prep, that technical work includes deployed agent harnesses, deployed model adaptations, and deployed evaluation-informed application configurations. I work on deciding what a model may propose, what application code must verify, and how to tell whether a change is actually useful. I also built and deployed a speech-distillation pipeline, training a Piper/VITS student voice on audio generated by a teacher model and evaluating exported candidates before release.

School Builder Bot began as a Virtu Prep project in 2021 and later became a separate project. Astral AI Academy began as a Virtu Prep project in 2023 and later became a separate project. Both grew from the work of building learning and training systems inside the company.

Alongside the software work, I have explored a related question: how can people demonstrate ability when institutions fail to recognize it? Merit Institute, the Real Merit Protocol, and Learning Without Gatekeepers each address that question in different ways.

The systems work and the related investigation rely on the same distinction: what can we observe directly, and what are we inferring about the learner? Learning Without Gatekeepers and the published Real Merit Protocol address that distinction. The same distinction also frames my developing questions about prediction error, timing, and adaptation.

I am investigating those questions while developing the scientific and experimental skills needed to test them rigorously, and while continuing to build learning systems. This work is separate from product claims. Its purpose is to formulate testable hypotheses, report any results, and revise them when evidence disagrees.

This is an independent research direction in development, with no institutional affiliation. It does not present a validated BCI, approved dissertation, or completed experimental result.