AI Development Projects in State of São PauloAI Development Projects in State of São PauloI’m taking part in the Lovable Challenge, and this time I decided to start with a real problem from a real person: my friend Ana Mi, an independent tattoo artist from São Paulo.
A lot of her bookings happen through social media. Clients ask about prices, sizes, availability, and payments and she has to answer everything manually.
So I thought: what if booking a tattoo felt more like visiting an art gallery?
That’s how INKROOM was born. 🖤
A place where people can discover Ana’s work, choose a tattoo, customize it, book an appointment, or even send their own idea for a custom quote.
And Ana can manage requests, appointments, payments, and her tattoo collection all in one place.
Less time managing messages. More time creating.
Built with Lovable for the Lovable Challenge 2026. Most OCR systems stop at text extraction.
The real value starts after the document is read.
In this project, I built an AI-powered OCR pipeline capable of extracting text from invoices, contracts, reports, and forms while also classifying the document type and identifying key information automatically.
Instead of returning raw text only, the system structures the data into usable outputs for automation workflows, CRMs, ERPs, and AI agents. This reduces manual processing time and makes documents immediately actionable inside business operations.
The stack combines OCR engines with LLM-based processing for contextual understanding, keyword extraction, and intelligent classification. The result is a more reliable document processing flow that can integrate directly with APIs, databases, and automation systems.
This is where AI engineering becomes practical: transforming unstructured files into structured business data at scale Hey Guys, new project Arived!
Montebello Água e Gás is a local water and gas distributor based in Saltinho, Brazil. The goal was to build a digital presence from scratch — fast, mobile-first, and conversion-focused. The project included target audience research and competitor benchmarking, which led to a deliberate choice: a lean, single-page layout that loads instantly and drives direct contact via WhatsApp. Delivered with Meta Pixel + Conversions API integration, Google Search Console setup, and SEO-optimized on-page structure — all built in HTML, CSS, JavaScript, Figma, Photoshop and Google Stitch deployed on a custom domain. THE CHALLENGE
Client relationships become difficult to manage when projects, conversations, account health, and revenue signals live in separate places.
THE SOLUTION
Clientio is a client management SaaS concept that gives teams one clear view of clients, active work, project health, timelines, and the next conversation.
WHAT I BUILT
• Client portfolio and health indicators
• Project status, revenue, and timeline views
• Central inbox and client activity feed
• Responsive management dashboard
• Clear information hierarchy for daily account work
DESIGN APPROACH
A bright blue workspace with warm orange highlights makes dense operational data approachable while preserving a confident professional feel.
THE OUTCOME
A polished CRM product demonstration combining product strategy, dashboard UX, visual design, and frontend engineering.
LIVE PRODUCT
https://clientio-saas-demo.almond-bowl-2897.chatgpt.site 
