SEP Dynamics is a foundational data intelligence firm commercializing the Self-Emergent Processor (SEP) Engine—a proprietary, high-performance C++ framework that redefines how complex data is analyzed by quantifying its informational coherence and stability.
Our claims are not theoretical; they are backed by a suite of working proofs of concept that we can demonstrate today.
The engine ingests any data as a raw byte stream, proven by its ability to produce distinct coherence scores for random data (0.0561), repetitive data (1.0000), and a complex compiled executable (0.4682).
Analysis of a large data stream is equivalent to the aggregated analysis of its parts, with a statistically negligible variance of less than 0.0015.
The engine can retain memory of past patterns for time-series analysis or be explicitly cleared for perfectly clean, reproducible backtests.
Built in modern C++ with a CUDA backend, the engine processes sample data in ~27 microseconds (~7.8 MB/s) and has proven linear scalability.
Multiple paths to participate in the quantum finance revolution
A fundamentally new approach to market analysis
B.S. Mechanical Engineering, University of Oklahoma (2019)
Alexander combines deep physics knowledge with proven execution in high-stakes engineering. From developing control systems for Mark Rober to mission-critical automation at Apple/Flex, he brings a unique perspective to quantitative finance.
The insight: Markets behave like complex thermodynamic systems. Traditional models fail because they assume equilibrium. The SEP Engine measures disequilibrium—the real source of alpha.
Finalize C++ engine implementation, optimize performance, and achieve a functional, compelling demo to validate the core thesis.
Transition to full-time dedication, establish SEP Dynamics LLC, validate 5 proofs of concept, and file for patents [docs].
Launch proprietary trading operations and begin generating revenue.
While others try to predict prices, we measure the quality of information in market data:
Markets are information processing systems. By measuring information quality directly—not derived statistics—we capture the true market state. Our proofs demonstrate:
The SEP GitHub repository referenced is currently private. Access may be granted to serious collaborators upon request.
All performance metrics and financial projections are illustrative only and do not guarantee future results.
Information about patent filings is provided for transparency. Four provisional patent applications have been filed and are currently pending approval. See our patent documentation.
Alexander J Nagy collaborated with Mark Rober on public projects such as Glitterbomb 2 and the July 2021 domino robot world record.