Financial systems | Agentic infrastructure | Quantitative research
I build financial systems where software can act on money, data, or infrastructure, and where authority, evidence, and financial semantics matter as much as model capability.
My background spans an MBA from NYU Stern, structured credit, financial-crime investigations, private-equity underwriting, and founder-led enterprise work. Before this current technical work, I founded and exited an enterprise compliance business, including an IP sale and a separate sale of the operating company.
Across the systems below, I work as founder and technical product lead. I set product direction, architecture constraints, acceptance criteria, and release decisions, and I review implementation at the source, test, and system-evidence level. AI-assisted engineering tools are part of the build workflow; technical direction and acceptance remain mine.
Controlled execution infrastructure for AI agents
BoxFetch controls how software agents acquire, authorize, and execute capabilities that can spend, mutate, or delete. The system combines MCP, OAuth 2.0 and PKCE, scoped permissions, purchase entitlements, spending controls, digest-bound human approval, idempotent state transitions, audit evidence, and a constrained execution runtime.
Focus: agent authorization, secure execution, transaction controls, human-in-the-loop decisions, PostgreSQL, TypeScript, Next.js.
Financial graph intelligence and probabilistic forecasting research
Monaith models accounts, transfers, repayments, recurring behavior, and cash flow as a connected temporal system. Its forecasting architecture separates deterministic known obligations from learned residual uncertainty and preserves a clear boundary between observed facts, model outputs, and product claims.
Focus: financial graphs, transaction reasoning, probabilistic forecasting, Plaid, PostgreSQL, TypeScript, Python model workers.
Scientific quantitative research infrastructure
SHARK is a short-horizon research laboratory designed to discover and falsify market effects under point-in-time information boundaries. It combines immutable evidence, canonical PostgreSQL memory, typed OCaml research, hostile methodological testing, and structured scientific memory that keeps machine results, scientist conclusions, and spent evidence distinct. Established quantitative finance, mathematics, statistics, and mathematical physics are treated as prior art and adversarial baselines; candidate combinations advance only when they add residual structure beyond simpler explanations.
Focus: OCaml, PostgreSQL, intraday market data, experimental design, scientific memory, numerical research, reproducibility.
Institution-oriented financial data infrastructure
PIASSES provides a provider-neutral architecture for consented financial data with REST and MCP interfaces over one application core. It emphasizes exact money representation, provenance, explicit absence semantics, durable consent state, idempotency, auditable authorization, and replaceable institution adapters.
Focus: financial data contracts, consent, provenance, MCP, REST, PostgreSQL, TypeScript.
- Financial semantics before convenience. A technically valid number is still wrong if its economic meaning is wrong.
- Authority must be explicit. Reading, spending, mutating, approving, and deleting are different powers and should be modeled that way.
- Evidence classes should remain distinct. Synthetic, sandbox, offline, live, and production evidence support different claims.
The underlying product and research repositories are private. These public case studies expose the architecture, decisions, validation evidence, corrections, and current system boundaries that can be shared safely.
The goal is to make the work inspectable without publishing proprietary code, customer data, credentials, or security-sensitive operational detail.