AI Automation Engineer

I build AI systems that run the work for you.

I design and ship autonomous pipelines — content generation, workflow automation and data processing — orchestrated with LLMs like Claude, GPT and Groq. From a single prompt to a published result, with no human in the loop.

View selected work → Get in touch
13-stage
autonomous pipeline
3× LLMs
Claude · GPT · Groq
0
manual steps
Selected work

Systems that do the job end‑to‑end.

Real automation projects — each one takes a messy manual process and turns it into a system that runs itself.

// 01 — FLAGSHIP

Auto Content Pipeline

End-to-end AI video automation · Node.js
Problem

Producing and publishing video is slow and manual — research, scripting, voiceover, editing, thumbnails, publishing and analytics all done by hand.

Solution

A pipeline that takes a topic and autonomously produces a finished video, then publishes it to YouTube and TikTok — fully unattended.

How it works

Thirteen chained stages orchestrating multiple LLMs for research and scripting, TTS voice, media assembly, AI thumbnails, multi-platform publishing and revenue tracking.

Result

A hands-off content operation: idea in, published video out — with built-in analytics and monetization tracking.

research▸ script▸ voice▸ media▸ thumbnail▸ video▸ publish▸ analytics
Node.jsAnthropic ClaudeOpenAI GPTTTSYouTube APITikTok APISharp
// 02

Content Research & Generation API

Serverless AI automation · Python
Problem

Manual research and drafting eats hours, and the output ends up scattered across notes and documents.

Solution

A serverless API that scrapes sources, processes them with AI and writes structured results straight into Google Sheets.

How it works

A Flask endpoint deployed on Vercel scrapes and cleans content, runs it through Groq and Claude, and syncs the output to a live spreadsheet the team already uses.

Result

On-demand research and content, callable as a single API request — no manual copy-paste.

PythonFlaskBeautifulSoupGroqAnthropic ClaudeGoogle Sheets APIVercel
// 03

Consistent AI Character Generation

Generative imagery pipeline · ComfyUI + LoRA
Problem

Generating a brand-consistent character or avatar across many images is hard — the identity drifts from one render to the next.

Solution

A custom image pipeline with a trained LoRA model that locks a character's face and body across unlimited scenes and poses.

How it works

Curated dataset and custom LoRA training on Stable Diffusion XL, driven through an automated ComfyUI workflow — including a from-scratch AMD/ZLUDA GPU setup.

Result

Unlimited on-brand imagery of a consistent character, from a reusable and documented pipeline.

ComfyUIStable Diffusion XLLoRA trainingPythonAMD / ZLUDA
// 04

Real-Time Market Monitoring & Alerts

Data pipeline & dashboards · Python
Problem

Watching markets and spotting signals by hand is impossible to keep up around the clock — the moment always passes before you see it.

Solution

An automated system that pulls live exchange data, runs technical analysis, visualizes it in a dashboard and pushes alerts straight to Telegram.

How it works

A Python service streams data from exchanges via ccxt, computes technical indicators with pandas, renders an interactive Streamlit + Plotly dashboard, and fires real-time alerts through a Telegram bot.

Result

Continuous, hands-off market monitoring with instant alerts — the data watches itself.

PythonccxtpandasStreamlitPlotlyTelegram Bot API
Stack

The toolkit behind the automations.

AI & LLMs

Claude APIOpenAI GPTGroqPrompt engineeringStable DiffusionLoRA training

Automation

n8nPipeline designWeb scrapingAPI integrationScheduling

Backend

Node.jsPythonFlaskPHPREST APIsJava

Web & Infra

TypeScriptReactJavaScriptVercelGitHTML / CSS
Arthur Paz About

Automation with a growth mindset.

I'm Arthur — a developer who builds AI automation aimed at a business result, not just a working script. I take a slow, manual process and turn it into a system that runs on its own: content pipelines, data workflows, and integrations between the tools a business already uses.

I come from a web development and growth background — SEO, retention, the whole user funnel — so I automate toward outcomes. Lately my focus is wiring LLMs like Claude, GPT and Groq into real, deployed pipelines, building the whole thing end to end: backend, integrations and the interface on top.

RoleAI Automation Engineer
FocusLLM pipelines
Based inBrazil · remote
Working withUSD / EUR clients
Status● Available
Contact

Have a process that should run itself?

Tell me what's eating your time — I'll tell you how to automate it.

Email me → LinkedIn GitHub WhatsApp