BUILDING WITH PROMPT ENGINEERING

Hey, I'm Salvador—a self-taught developer who's spent months learning to build through prompts.

Not talking about it. Actually building it:
Production apps that work, deployed and live.

No team. No funding. Just me, AI models, and a lot of iteration.

Here's what I've built.

HOW I GOT HERE

I'm self-taught. No CS degree, no bootcamp.

What I do have? Months of experimentation with prompt engineering.

I started because I was curious if you could really build production apps by orchestrating AI models.

Turns out you can. But it takes work:

  • Learning which prompts work (and which don't)
  • Teaching Claude, ChatGPT, and others to build real systems
  • Debugging when things break (they break a lot)
  • Iterating until it actually works
  • No shortcuts—just focused effort

The learning curve was steep. I learned to:

  • Write prompts that generate Python backends with proper architecture
  • Get Rust code that actually compiles (not tutorial examples)
  • Design PostgreSQL schemas through conversation
  • Build production systems, not demos

Most days were frustrating. Code broke. Models hallucinated. Nothing worked.

But I kept iterating.

And now I have three working apps built entirely through prompt engineering.
Deployed. Live. Actually working.

This is what's possible when you put in the work.

THREE APPS BUILT WITH PROMPTS

Not demos. Not prototypes. Real applications you can use today.

🧑‍💻

TIO - Your Coding Uncle

"Everybody wishes they had a coding Tio"

The Replit killer I built out of spite (and necessity).

What It Does:

  • Generates production-ready code in Python, Rust, TypeScript, and more
  • Actually handles complex stuff (PostgreSQL, proper architecture)
  • Deploys to Railway, Render, Fly.io
  • Fixes broken code automatically
  • Not a toy - this is the real deal

Why It Matters:

Replit charges $1000/week and can't even do Rust properly.
Tio can. For way less.

✅ Just Deployed 🚀 Users: Early access available
📈

ARBITRADE - AI Trading Signals

AI-powered market analysis

What It Does:

  • AI-powered trading signal analysis
  • Arbitrage opportunity detection
  • Real-time market analysis
  • Helps traders find edges

Why It Matters:

If AI can generate code, it can analyze markets.
Same principles, different application.

Launching Soon Beta access available
🌺

MELERY - Mints & Manages Solana Tokens

Launch a Solana token in 30 seconds, manage it forever

What It Does:

  • Create professional SPL tokens instantly
  • Auto-create Raydium liquidity pools
  • Lock LP tokens (native Solana, no third-party)
  • Dashboard to manage & withdraw locked LP
  • Fully automated—no technical knowledge needed
  • 0.5 SOL + 1% allocation OR 0.6 SOL + 0%

Why It Exists:

Token launches shouldn't require a dev team.
MELERY mints your token, locks your LP, and gives you a dashboard to manage everything. No scrambling to find your locked tokens later.

✅ Live on Solana Zero support, fully automated

WHAT PROMPT ENGINEERING TAUGHT ME

I'm not employed by OpenAI, Anthropic, or Meta.
I'm just someone who spent months learning how to build through prompts.

Here's what I've figured out:

  • How to make AI write systems-level code (Rust, proper Python)
  • How to design prompts for complex database architecture
  • How to get production-quality output, not tutorial code
  • Which models are good at what (they're not all the same)
  • How to orchestrate multiple AI tools together
  • What works, what breaks, and why

I've spent more time on this than most people.
Not because I'm special—because I was curious enough to keep trying.

3
APPS
Built with prompts
100%
AI-BUILT
No traditional dev
LIVE
DEPLOYED
Actually working

HOW THESE MIGHT HELP YOU

Need code generation?

Tio handles multi-language projects, deploys to production platforms, and fixes broken code.

Try Tio →

Launching a Solana token?

MELERY creates your token, sets up the pool, and locks LP—all in 30 seconds.

Launch with MELERY →

Trading crypto?

Arbitrade analyzes markets and finds arbitrage opportunities using AI.

Visit Arbitrade →

Learning prompt engineering?

I share what I learn on TikTok—the wins, failures, and what actually works.

Follow on TikTok →

BEING HONEST ABOUT WHERE I'M AT

What I'm Not:

  • A consultancy with a team
  • An established company with paying customers
  • Available for big projects right now
  • Pretending to be bigger than I am

What I Am:

  • A self-taught developer learning through prompts
  • Someone who built three working apps this way
  • Building tools you can actually use
  • Sharing what I learn publicly
  • Open to interesting opportunities

THREE WAYS TO GET INVOLVED

🚀

Use What I Built

See if these tools solve problems for you.

📱

Follow What I'm Learning

TikTok: @salvador_the_brown

I share what works (and what doesn't) in prompt engineering.

Follow Now
💼

Work With Me

Once these apps are running smoothly, I might take on consulting work.

Right now I'm focused on building.

Email Me

TRY WHAT I BUILT

See if it works for you.

Three apps. All built through prompt engineering.
All deployed and working.

You can:
→ Use the tools I made
→ Follow what I learn
→ See where this goes

Or check back in a few months and see what's changed.

Either way, this is what's possible when you combine curiosity with prompts.