Preface
This volume documents a series of investigations into computational systems, conducted between 2020 and the present day. The works catalogued herein span from low-level systems programming to high-level application development, with particular attention to artificial intelligence, distributed systems, and user-facing software.
The specimens were collected across various ecosystems: open-source repositories, production environments, and academic institutions. Where possible, I have preserved the natural habitat information and technical classifications for future researchers.
250+
Students Taught
20+
Colleges
10
Public Repos
4
Core Languages
The reader will note a pattern: each work represents an attempt to solve a problem that probably should not exist in the first place. Nevertheless, they persist.
Plates I–X: Specimen Collection
Plate I
ArnoldC / ToaruOS
Classification: C, Assembly, Systems Programming
Habitat: GitHub, various open-source repositories
Field Notes: Built a complete desktop OS in ArnoldC — 22,000 lines where every keyword is an Arnold Schwarzenegger quote. GUI, networking, games, 213KB binary.
Yes, really.
Plate II
Erdos
Classification: Python, Lean 4, Multi-agent LLM
Habitat: GitHub
Field Notes: Automated theorem prover — Prover/Critic multi-agent loop with Lean 4 and SHA-256 integrity locks. The agents kept redefining what they were proving, so I had to lock the theorems down.
Erdős would approve. Probably.
Plate II-B
polymarket_bot
Classification: Python, Kelly criterion, Monte Carlo
Habitat: GitHub
Field Notes: Prediction market trading with regime detection and Kelly criterion sizing. Knows when not to trade.
The trap detector is the best part.
Plate III
jobhunter
Classification: Python, TypeScript, Gemini AI, GCP
Habitat: GitHub
Field Notes: AI job application system that scrapes, scores, and drafts — but won't submit without human approval and won't lie about credentials. 52 applications drafted, 0 auto-submitted.
The honesty constraints are the interesting part.
Plate III
ai-tax-cpa
Classification: Python, LLMs, RAG
Habitat: GitHub
Field Notes: AI tax assistant that actually understands Schedule C.
The IRS remains unimpressed.
Plate IV
Bee
Classification: React Native, Expo, TypeScript, tRPC.js
Habitat: GitHub
Field Notes: Spelling Bee word game — hexagonal letter grid, 370k-word dictionary, 5 hint types, 4 difficulty modes. WCAG AAA accessible. 80 tests.
Turns out humans need help organizing.
Plate V
HealthOS
Classification: Python, FastAPI, React, Docker, JWT
Habitat: GitHub
Field Notes: Health tracking platform — medical records, lab results, medications on one timeline. FastAPI + React, JWT auth, Docker.
Your Apple Watch data, but useful.
Plate VI
claude-sfx
Classification: Node.js
Habitat: GitHub, npm registry
Field Notes: Sound effects integration for Claude Code. Makes your terminal less quiet.
Silence was getting awkward.
Plate VII
anti-slop-design
Classification: Design Methodology
Habitat: GitHub
Field Notes: Design system that kills generic AI aesthetics. 8 domain profiles, 15-rule checklist.
Someone had to say it.
Plate VIII
saas-starter-kit
Classification: Next.js, Stripe, PostgreSQL
Habitat: GitHub
Field Notes: Production-ready SaaS boilerplate. Auth, billing, dashboards.
So you can skip the boring parts.
Plate IX
BennyCuTools
Classification: TypeScript, Cloud Run, GCP
Habitat: GitHub, production deployment
Field Notes: Personal infrastructure toolkit. Agents, dashboards, automation.
Personal infrastructure is infrastructure.
Plate X
cuuper22.github.io
Classification: HTML, CSS, JavaScript
Habitat: GitHub Pages
Field Notes: This portfolio (well, one of ten versions of it).
You're looking at variant 10 of 10.