Intentionally overengineered personal finance tracker (Go + React) with microservices, immutable ledger, full CI/CD, and extensive documentation. Well-architected but zero external adoption or users.

Turorepo monorepo dashboard for observing pi coding agents with typed architecture, full CI/test coverage, SSE real-time updates, and well-structured 5-package design (server, client, desktop, extension, shared).

Polished personal knowledge capture app: macOS Swift UI + Python FastAPI backend with hybrid search, LLM-powered tagging, and pgvector embeddings. Well-documented, CI/CD configured, structured multi-service architecture. Early-stage indie project with working implementation.

Sophisticated 3D bin packing solver with tiered algorithm (volume filter → heuristic → HiGHS MIP), full test coverage, Lean 4 formal verification, and interactive React+Three.js UI. Launched yesterday with 13 commits; no stars yet.

Modular-monolith learning-roadmap platform with FastAPI backend, React frontend, and MCP server. Typed Python+TS, comprehensive CI/tests (80% backend coverage, 70% frontend), multi-layer architecture with OAuth 2.1, but brand new (7 days old, 30 commits) with zero external adoption.

Floresu is a polished career tracker with AI agent integration. Well-structured Python/TypeScript full-stack project with strict types, comprehensive CI (6 jobs), tests, and detailed docs covering OAuth2, embedding, search, and resumes—but remains a personal project launched 14 days ago with 0 stars.

TypeScript Obsidian plugin wrapping Excalidraw with minimal UI overhead. Ships typed code, tests, autosave codec, and comprehensive markdown serialization. New project (1 week old) with clean architecture but zero adoption signals.

Personal portfolio site built with SvelteKit + TypeScript featuring a procedural pipes background animation (canvas rendering + WebGL CRT shader) with full test coverage, CI/CD automation, and live deployment. Well-structured codebase with meaningful project documentation and typed components.

Early-stage scheduling system with comprehensive docs and production infrastructure (CI/CD, tests, Docker compose) but zero adoption. ~8.4 MB codebase demonstrates architectural ambition across Python backend (CLI, domain logic, learning), TypeScript frontend, and E2E testing.

Single-purpose GitHub Actions monitoring script for Hetzner server availability with clean design and standard library only, but brand new (3 commits over ~11 minutes), no tests, and minimal scope.