Start Building — Elitez Engineering

GETTING STARTED

Five steps, not five weeks.

STEP 1

Pick one small annoyance

Not a grand vision — one repetitive task you personally do too often. Browse Ideas Wanted on the homepage if nothing comes to mind.

STEP 2

Get access to an AI tool

Use Claude or our local LLM (e.g. Qwen Code) — most projects here started with just a chat window and a plain-English description of the problem.

STEP 3

Describe what you want, in plain English

You do not need to know how to code. You need to be able to explain a problem clearly — the same skill you already use to brief a colleague.

STEP 4

Ask when you're stuck

Ask on WhatsApp — nobody expects you to figure everything out alone.

STEP 5

Ship something small, then tell us

A working prototype beats a perfect plan. Once you have anything real, use the form below — we will get you a card on the portfolio and on the builders page.

GET SET UP

I want to try this.

Fills in an email so we know you're interested — this is not wired to a shared tracker yet, so a real reply will follow.

FAQ

The real objections.

"I'm not technical."

Neither is anyone on this list. The skill that matters is describing a problem clearly, not writing code.

"I don't have time."

Most projects here were built in evenings alongside a full-time day job — see the commit counts on each project card. Start with something that takes an hour, not a quarter.

"What if I break something?"

Start on your own idea, on your own time, separate from production systems. Nobody's first attempt touched a live system that mattered.

"Will this replace jobs — including mine?"

Every project here was built to remove repetitive work, not people — freed-up time has gone to judgment calls and higher-value work, not headcount cuts.

"Do I need permission?"

No. We expect creativity and taking initiative — see also the note from leadership on the journey page.

"What tools do I actually need?"

Claude or our local LLM (Qwen Code) — most of what's live here needed nothing more than a chat-based AI tool to start.

GLOSSARY

Jargon, decoded.

Every term you will see on a project card, in plain English. Hover any tag on a project card for the same definitions inline.

RAG

Retrieval-Augmented Generation — an AI looks up your real documents first, then answers using what it found, instead of guessing from memory.

LLM

Large Language Model — the general engine (like Claude or Qwen) behind most AI text/chat features.

Local LLM

An AI model running on our own hardware instead of a cloud API — private, and free to run once set up.

Ollama

Free software that runs AI language models directly on a laptop, no internet or cloud account needed.

On-device AI

AI that runs directly on a phone/laptop/Mac instead of calling out to the internet — faster and free per use.

YOLO

A fast, well-known AI model family for spotting/locating objects in a photo (e.g. bottles on a shelf).

VLM

Vision-Language Model — AI that can look at an image and describe or answer questions about it in plain text.

Multi-agent AI

Several AI workers, each handling one sub-task, coordinated together to finish a bigger job.

Embeddings

A way of turning text into numbers so a computer can find similar-meaning content, not just keyword matches.

Supabase

A free/low-cost hosted database + login system — a common starting point for a real web app.

API

Application Programming Interface — a way for two pieces of software to talk to each other automatically.

Cloudflare Access

A free login gate that restricts a website to only people with an approved company email.

Repo / repository

The folder of code for a project, usually tracked with version history (see commit).

Commit

A saved checkpoint of code changes, with a timestamp — how we know when and how much work happened.

Deploy / deployment

Publishing a build so it is live and usable, not just sitting on your own laptop.

MVP

Minimum Viable Product — the smallest version of an idea that is still genuinely useful.