Labs · Engineering & Exploration

Engineering & Exploration.

A curated collection of open-source solutions and technical explorations.

This is a space for innovation, where architectural strategy meets hands-on implementation. From micro-conditional engines to legacy data parsers, each project is a battle-tested solution born from a passion for solving complex software challenges.

01The Illogical Ecosystem
02The Nexus Grid
03Nexus Labs: The Assist
04Project: Space Wars
05Architectural Tooling
06Project: Pac-Man

01 / The Flagship

The Illogical Ecosystem

A micro conditional engine for parsing and evaluating raw logical expressions, built to solve the recurring problem of sharing complex business logic between clients, services, and edge functions without redundant implementations.

A glowing amber logic gate splitting a beam of light into branching paths

One engine, seven runtimes

A micro conditional engine used to parse raw logical and comparison expressions. It evaluates expressions in a given data context, ensuring consistent logic across your entire stack.

Implementations

rule.jsonIllustrative

// one rule, shared as data
["AND",
  ["==", "$plan", "pro"],
  [">=", "$seats", 10],
  ["IN", "$region", ["CA", "US"]]
]

// against a context
{
  "plan": "pro",
  "seats": 12,
  "region": "CA"
}

Same verdict, every runtime

TypeScripttrue
Pythontrue
Gotrue
C#true
Javatrue
Odintrue
Zigtrue

02 / Platform Engineering

The Nexus Grid

Internal developer sandbox fostering engineering collaboration. A polyglot microservices platform where teams build utilities for peers, experiment with AI tooling, and create an evergreen onboarding playground.

An isometric grid of modular blocks with amber connections lighting up between them

Nexus Grid Platform

Internal Developer Sandbox • Tactable Engineering

Modular utility platform for engineering team. Bring Your Own Language philosophy enables polyglot microservices. Each engineer builds tools peers use.

Serverless backend functions (GCP Cloud Functions) + React frontend. Auto-deploy via Terraform. Teams work parallel, zero conflicts.

⚡ GCP Cloud Functions🏗️ Terraform IaC⚙️ GitHub Actions CI/CD🔐 Google SSO Auth

Core Objectives

Build Useful Utilities:Unified platform for small tools. Daily workflow automation, custom processors, internal aggregators.
Master AI Tooling: Active lab for AI-assisted dev. GitHub Copilot experimentation in real-world collaborative setting.
Evergreen Sandbox: Ultimate onboarding playground. Safe environment for new hires to commit code, learn CI/CD, see utilities go live.

Tech Stack

FrontendReact 19 + Vite

UI Frameworkshadcn/ui

BackendCloud Functions

InfrastructureTerraform

Package ManagerBun

LinterBiome

Platform Stats

25+

Active Utilities

Language Runtimes

100%

Serverless

∞

Scalability

Modular Architecture

Backend Functions

Independent micro-utilities. Any language. Isolated functions.

Node.jsPythonGoJavaRuby.NETPHP

Proxy API

Central nervous system. Function registration. Route handling.

Cloud Run

Frontend Widgets

Self-contained UIs. Plug into central dashboard. React components.

React + Vite

Engineering Impact

Zero

Deployment Bottlenecks

Config-driven auto-deployment. Teams work parallel.

∞

Language Freedom

Node.js, Python, Go, Java, PHP, .NET, Ruby support.

100%

AI-Assisted

Active GitHub Copilot experimentation lab.

03 / Engineering Culture

Nexus Labs: The Assist

A two-day sprint where sixteen engineers each built a tool or agentic system to pay down their own daily tax: the repetitive triage, status updates and manual toil that quietly eat a week.

Sixteen glowing workstations in a circle, beaming into one shared point of light

Nexus Labs: The Assist

2-Day Engineering Sprint • Tactable Engineering

Every engineer pays a personal tax: the same triage, the same status update, the same boilerplate PR comment, the same "let me dig up that log again."

The Assist is a lab where each engineer builds a tool or agentic system that pays down their own daily tax. No client constraints, no team-wide project to coordinate, just a real annoyance from your week and the AI tooling already on your machine.

🧑‍💻 Solo Build🎤 Peer Showcase🤖 Copilot & Claude📦 Shared Cohort Repo

The Brief

Make It Real: Target friction in your actual daily work: a repetitive task, a request you keep fielding, recurring manual toil. It doesn't need to be big; it needs to be real.
Ship Tooling, Not Prompts: Build actual tooling or an agentic setup, not a prompt you ran once. A working demo on Day 2, never just slides.
Leave It Behind: Bonus points when the build serves peers too: a shared script, a reusable agent config, a small internal tool committed to the cohort repo.

Quick Facts

Duration2 Days

FormatSolo + Demo

Cohort16 Engineers

Slot Length12–15 min

ToolingCopilot / Claude

Lab Stats

Days End-to-End

Parallel Builds

Tooling Tracks

Shared Repo

The Two-Day Format

Day 1 · Build

Solo build

09:30Kickoff: brief, scope, expectations, where to get help, then go build.
10:00Heads-down individual build time. No scheduled syncs for the rest of the day.
EODProgress, blockers and wins posted to the lab channel. Pairing encouraged, never required.

Day 2 · Showcase

Full team

~8 minDefine the problem, walk through the solution, live demo.
~5 minQ&A from peers, hard cutoff to keep the day honest.
ScaleSingle track fits ~20 people into one day; two parallel rooms halve the runtime as the cohort grows.

What Came Out of It

Sixteen independent builds, each aimed at a different personal tax. Clear clusters emerged without anyone coordinating them.

Knowledge Retrieval

Agents that answer "how does X work" over Confluence, project docs and system internals, so the same question stops landing in the same inbox.

Diagnosis Assistants

Stack-trace and deployment-failure explainers that surface a root cause and a recommended fix, grounded in a library of bugs already solved.

Status Synthesis

Weekly digests and meeting-note generators stitched from commits, tickets and transcripts, reconstruction work nobody should do by hand.

Review & Code Hygiene

Standards-aware review skills and a triage CLI that ranks stale TODO/FIXME markers by relevance and severity instead of dumping all of them.

Context & Memory

Ticket-scoped working memory over MCP: logging decisions, rejected alternatives and reversals so the reasoning survives to PR time.

Scaffolding & Publishing

Template repos for one-day POCs, plug-and-play project knowledge bases, and a skill that pushes Markdown straight from the editor to Confluence.

What the Format Gets Right

Own

Problem, Own Motivation

Scratching your own itch removes the need to sell anyone on the brief.

48h

Forces Real Scope

A demo deadline two days out kills gold-plating before it starts.

16×

Tooling Reps at Once

The showcase spreads sixteen engineers' hard-won AI lessons in one sitting.

04 / AI & Simulation

Project: Space Wars

An interstellar arena where engineers program AI bots to pilot spaceships, competing in a battle of wits, strategy, and coding prowess.

Small ships with amber engine trails dueling through a dark nebula

Space Wars was conceived as a team-building exercise and internal coding jam designed to fortify our collective passion for technology. It's an interstellar arena where engineers program AI bots to pilot spaceships, competing in a battle of wits, strategy, and coding prowess. More than a game, it's a laboratory for rapid iteration, collaborative problem-solving, and celebrating the joy of engineering.

View Project on GitHub

Technical & Strategic Highlights

Polyglot & WASM Integration

Engineers implement SpaceshipManager interface in any language that compiles to WASM (or use pure JS/TS), fostering a polyglot environment.

Complex Decision-Making

Bots must manage a dynamic arsenal of engines, lasers, and rockets, requiring complex, real-time strategic programming.

Physics-Based Simulation

The arena includes environmental hazards like asteroids, demanding skillful pathfinding and collision avoidance algorithms.

Rapid Iteration Loop

The game loop provides immediate feedback, encouraging engineers to quickly analyze, tweak, and redeploy their strategies.

Competitive Collaboration

While competitive, the format encourages knowledge sharing as teams observe and learn from the success of different strategies.

A catalyst for fortifying the collective passion for technology.

05 / The Utility Belt

Architectural Tooling

Specialized libraries for specific architectural challenges. Each tool solves a distinct problem in modern software development.

Data IntegrityTypeScript

WeGood

Shared validation logic designed to ensure data integrity across distributed systems. It creates a single source of truth for validation rules shared between UI and API.

View on GitHub

Search AlgorithmsTypeScript

Librarian

A micro search library leveraging Boolean Operators. It supports wildcard annotation within search terms, enabling powerful in-memory search for lightweight applications.

View on GitHub

Legacy IntegrationPython

X12 Parser

A robust parser for the X12 EDI standard. It handles the complexity of X12 loops and segment-based schemas, bridging the gap between modern Python data structures and legacy financial/health data.

View on GitHub

Developer ExperienceGo

ApiDoc

A generator that transforms Go source files directly into OpenAPI v3.0.2 (Swagger) specifications. Keeps documentation strictly in sync with code implementation.

View on GitHub

Frontend RoutingSvelte

Svelte Router

A declarative, component-based router for Svelte, inspired by React Router. It provides a simple and powerful way to handle routing in Svelte applications.

View on GitHub

Frontend ToolingSvelte

Svelte Form

A lightweight model handler for Svelte, streamlining input validation and state management for complex forms.

View on GitHub

06 / Digital Archaeology

Project: Pac-Man

A recreation of a classic, built to inspire and educate. A testament to the timeless engineering principles that underpin the craft of software.

A dark arcade maze outlined in glowing amber neon with dots receding into the distance

A deep dive into the heart of a classic. This implementation of the 1980s Pac-Man was an educational exploration, meticulously recreating the iconic game mechanics and ghost behaviors from publicly available materials. It stands as a tribute to the timeless engineering principles found in the jewels of early software development, a source of inspiration for aspiring engineers to see how a passion for the craft can translate into tangible, interactive history.

Explore the Archaeology on GitHub

07 / The Craft

A Commitment to the Craft

Each project represents a commitment to solving real-world architectural challenges. From high-level abstractions to systems programming, this work showcases how strategic thinking translates into production-ready solutions across diverse technological contexts.

Programming Languages: 7+
Architectural Tools: 5
Cross-Platform Standard: 1
Problems Solved: ∞