Software engineer / banking .NET apps / product tools

Maxwell Metzner

I build and support enterprise banking software with C#/.NET, WPF/WinForms, SQL Server, and REST/SOAP service layers at Fiserv. Outside that work, I ship browser extensions, desktop utilities, solver engines, and static web apps that turn messy workflows into usable tools.

View Projects Resume

Banking .NETEnterprise teller/core-banking apps, release support, SQL Server, WPF
AutomationPython and .NET utilities for logs, test data, SharePoint/forms, and imports
Product systemsStatic demos, OAuth PKCE, browser workers, solvers, and data workflows

01 Explore Selected Work 02 See Site Architecture 03 Get In Touch

Selected work

Projects with visible mechanics, not just labels.

The featured work below points to running demos or source and calls out the actual pieces worth inspecting: solver files, worker pipelines, auth flows, desktop media paths, ranking logic, and UI surfaces.

Engines

Exact EV tables, backtracking solvers, route scoring, media analysis, and browser workers are visible in code.

Interfaces

The previews mirror the apps' real layouts: dashboards, puzzle grids, dice tools, tables, and timelines.

Constraints

Static hosting, PKCE auth, local storage, client workers, and desktop dependencies shape the implementations.

Proof

Demo and repo buttons are kept close to each project so the claims are easy to check.

01 / Analysis Engines

Casino Cheat Sheet

Static Next.js dashboard with separate analyzers for roulette, baccarat, craps, blackjack, video poker, pai gow, and poker equity.

Splits a pure TypeScript casino engine package from the web UI, then pushes slower equity and simulation work through a browser worker. Visitors can test combined roulette bets, finite-shoe blackjack EV, video-poker holds, and multiway Hold'em/Omaha/stud without any server runtime.

The important files are easy to inspect: packages/casino-engine/src owns the math, while apps/web/src/workers/casino-worker.ts keeps the static UI responsive.

Live pages: 10
Worker: casino-worker
Engine: TS package

Demo Code

Casino Cheat Sheet Exact analyzers

Finite-shoe blackjack, poker equity, and rule-driven table games.

BlackjackEV solve Video pokerHold rank Hold'emEquity CrapsTable state

browser worker queue

02 / Algorithms

LinkedIn Puzzle Solver

Plain HTML/CSS/JS solver surface for Zip, Tango, Queens, Patches, and Mini Sudoku.

Each puzzle has its own grid/input layer and solver file under js/solvers. The app validates impossible boards before search, then marks clues, generated cells, conflicts, and solved paths in the same interface used to enter the puzzle.

The value is visible in the UI: tabs, board editing, solve/reset controls, and solver-specific feedback rather than a black-box answer.

Solver files: 5
UI: Plain JS
Validation: Before search

Demo Code

03 / Game Theory

Farkle Lab

Playable human-vs-CPU Farkle with exact EV tables and an advisor panel.

Precomputes expected values for one through six remaining dice, highlights legal scoring choices after each roll, and compares roll-vs-bank decisions against score, target, and late-game pressure.

Open app.js and the demo side by side: the game loop, scoring choices, CPU decisions, and advisor all share the same EV logic.

EV tables: 1-6 dice
Opponent: CPU
Advisor: Roll/bank

Demo Code

04 / Data Workflow

Spotify Manager

Dark, table-first playlist workspace with browser-only Spotify PKCE auth.

After setup, the app loads playlists, supports CSV-enriched metadata, sortable/resizable Tabulator columns, dedupe review, outlier scoring, transition diagnostics, undo/redo, and export to a new Spotify playlist.

It shows real product constraints in static hosting: OAuth callback pages, local token/config storage, and a dense data-table workflow.

Auth: PKCE
Table: Tabulator
History: Undo/redo

Demo Code

TitleBPMEnergy Night Drive1240.82 Arcade Bloom1180.74 Static Heart1320.89

Mix AssistTransition diagnostics

05 / Media Tools

Video Stabbot

Electron video stabilization UI that wraps FFmpeg vidstab, OpenCV features, and RAFT optical flow.

Detects dependencies and GPU encoders, accepts drag-and-drop video input, offers preset and advanced stabilization modes, streams progress, and calls Python/OpenCV/PyTorch scripts for heavier analysis paths.

This is a desktop systems sample: renderer UX, main-process detection, external binaries, Python workers, and media pipelines have to cooperate.

Input: Drag/drop
Engines: 3 paths
Encoder: GPU detect

Code

06 / Planning Systems

Trip Planner

React/Vite itinerary planner that scores stops against route timing, detour, daylight, and preferences.

The guided form builds a route request, then the planner slots meals and optional stops into a timeline. It can run in deterministic demo mode or browser-side Google mode, with local saved trips and explanations for each recommendation.

The ranking pipeline is visible in src/lib/planner: slots, scoring, daylight, provider adapters, and replace/pin/skip interactions stay client-side.

Planner: Client-side
Modes: Demo/Google
Explain: Score reasons

The profile README and GitHub profile configuration repository.

Repository config Mar 2026

Code

Site architecture

Static runtime, data-driven maintenance, progressive enhancement.

The page is committed as semantic HTML for indexing and no-JS access. Project content is generated from JSON so adding, removing, and reordering work stays boring in the best possible way.

Canonical project data

assets/data/projects.json owns project order, metadata, links, featured status, and preview type.

Generated static markup

npm run build:projects writes SEO-friendly project HTML and JSON-LD into index.html.

Small enhancement layer

Vanilla JavaScript adds filtering, sorting, reveal animation, theme persistence, and a reduced-motion-aware canvas field.

Performance first

No framework payload, no external fonts, lazy media below the fold, CSS containment, and static assets suitable for GitHub Pages.

About

Engineer at Fiserv, builder after hours.

I am a software engineer based in New Jersey with a Computer Science degree from Stevens Institute of Technology. Professionally, I work on Fiserv enterprise teller and core-banking software with C#, .NET, WPF/WinForms, SQL Server, and service-layer integrations.

Outside of work, I like building practical tools with enough depth to teach me something: browser automation, static apps that handle real auth flows, Windows utilities, and algorithmic engines that stay understandable.

C# .NET Python TypeScript JavaScript WPF SQL Server Chrome APIs GitHub Pages

Contact

Have a project, role, or strange workflow worth untangling?

I am open to software engineering opportunities and collaborations around developer tools, automation, static web apps, and practical product engineering.

Email GitHub LinkedIn Resume