Independent thinking. Practical software.

Building tools
for complex
questions.

Lee · Independent Software & Research Developer

Exploring numerical methods, engineering reliable systems, and making technical knowledge easier to navigate.

Explore the work A little about me

Ideas, made tangible

01 / Scientific computing 02 / Trading infrastructure 03 / Knowledge tools

Selected work

Three areas of exploration.

Research code, operational systems, and focused tools for working with complex information.

01 / Numerical methodsλ

Boundary conditions → collocation → solution
Illustrative method sketch

Scientific computing / Research software

FR-PIELM &
Particle Simulation

A research codebase connecting neural-network numerical methods with one-dimensional electrostatic simulation, alongside a focused reference implementation of the Poisson solver.

FR-PIELM and finite-difference Poisson solvers with embedded boundary conditions in the neural formulation.
Particle-in-cell routines for charge deposition, field interpolation, and electron–ion dynamics.
Reference solver contracts, rank and condition checks, and numerical diagnostics.

Python
NumPy
SciPy
FR-PIELM
PIC

Inside the method

The solver constructs an analytic neural representation with hard Dirichlet boundary embedding. Its implementation uses matrix rank and conditioning checks, plus LU or QR factorization for the collocation system. The surrounding simulation code includes finite-difference comparison paths and particle updates.

The reference work makes assumptions and diagnostics explicit. Broader convergence guarantees and trustworthy coupling to particle simulations remain research questions.

Scope: one-dimensional numerical methods and implementation analysis.

02 / System engineering↳

Evaluate → control → record
Illustrative system sketch

Quantitative tooling / System reliability

Trading Research &
Execution Infrastructure

A Python system for quantitative research and trading workflows, with cost-aware paper execution, explicit risk controls, and a persistent record of system decisions.

Paper fills with fees, slippage, balance checks, and duplicate-request handling.
Allocation and loss limits, spread filters, cooldowns, and circuit-breaker checks.
Immutable experiment definitions, integrity-checked artifacts, and transactional journals.

Python
SQLite
HTTPX
WebSockets
PowerShell

Inside the system

Experiment definitions bind code, data, protocols, and selection rules. Append-only registration and content-addressed artifacts help keep research evidence traceable. Persistent journals capture orders, fills, positions, signals, and equity observations.

The engineering focus is explicit state, bounded risk, and inspectable execution behavior.

Experimental research and engineering; no claim of proven profitability.

03 / Knowledge interfaces≡

Structure → navigate → understand
Illustrative interface sketch

Document engineering / Offline-first tools

Technical
Offline Reader

A document-processing workflow and static desktop reader for navigating technical standards, with bilingual reading, offline assets, and source-aware content structures.

Semantic tables of contents, full-text search, glossary navigation, and bookmarks.
Offline math rendering, structured figures and tables, and source comparison views.
A content compiler and quality gates for structure, translation review, and visual evidence.

Python
JavaScript
HTML / CSS
KaTeX

Inside the workflow

Source material is organized into an intermediate content representation, then compiled into a static reader. The workflow keeps document structure, reading units, references, figures, and formulas tied to their source.

Bilingual reading connects a translated text with its source for side-by-side comparison. Quality checks cover content coverage, review evidence, asset integrity, and rendering. The generated reader opens locally without a web server.

Designed for desktop reading of technical standards.

About

A research mindset.
A builder’s approach.

I'm Lee, an independent software and research developer. My work spans numerical simulation, trading-system engineering, and tools for reading complex technical material.

Across these projects, I focus on making methods explicit, system behavior inspectable, and technical information easier to work with. I’m interested in the space where mathematical ideas become useful, understandable software.

Technical focus

What connects the work.

Three complementary directions, grounded in the projects above.

01 / MODEL

Scientific computing

Numerical linear algebra, boundary-value problems, neural numerical methods, and particle simulation. An emphasis on assumptions, diagnostics, and method clarity.

Python · NumPy · SciPy
Poisson solvers · PIC · Numerical analysis

02 / ENGINEER

Inspectable systems

Persistent state, controlled execution, risk checks, and traceable experiment artifacts. Software designed to make its decisions and operating behavior visible.

Python · SQLite · HTTPX · WebSockets
Event journals · Risk controls · Research tooling

03 / CLARIFY

Knowledge interfaces

Structured document pipelines, offline reading experiences, and source-aware navigation. Preserving the relationships between text, equations, figures, and references.

HTML · CSS · JavaScript · KaTeX
Content modeling · Search · Offline interfaces

Contact

Let’s talk about the work.

For questions about these projects or a technical discussion, get in touch.

aaagoooodman@gmail.com

Understand the method. Make it useful.

Building toolsfor complexquestions.