My work combines building software from start to finish with cool machine learning applications. Here are a few projects that show what I can do:

• LimpeJá is a mobile-first marketplace I founded and fully developed as the sole engineer. It connects clients to cleaning professionals with the same practicality of Uber and trust layer of Airbnb. The project was designed end-to-end by me -- from product vision and market analysis to mobile app, backend architecture, and deployment in the cloud.

• Core Features:

  • 📱 Clean and intuitive UX (Uber-like): ultra-fast flow from splash → login → signup → personalized home → booking → PIX payment. Includes smart recommendations, cancellation policies, and in-app chat for seamless communication.
  • 👩‍🔧 Empowerment for providers: complete autonomy to set prices, choose services, and manage schedules. Verified onboarding with document upload + facial recognition. Providers gain professional dashboards with history, earnings summary, and withdrawal requests -- elevating the experience to "driver-of-app" status.
  • 💸 Financial transparency: real-time earnings dashboard, upcoming jobs, and reviews. Instant PIX withdrawals (<24h) eliminate the #1 pain in the informal sector: uncertainty of payment. Dynamic pricing and surge incentives help fill idle hours and increase income.
  • 🔐 Security-first: robust KYC (OCR + selfie + background check), device fingerprinting, and fraud prevention for coupons. LGPD compliance highlighted in onboarding for client trust. Optional service guarantee/insurance with simplified claims process directly in-app.
  • 🤝 Reputation system: two-way ratings (client ↔ provider), badges like "Top of the neighborhood" and "100 services 5⭐", plus local ranking boards to reward consistency and quality. This builds community-driven trust.
  • 📊 Growth mechanics: referral program with attractive credit, first-service coupons, recurring bookings with discounts, and subscription plans that include insurance and priority scheduling. Marketing playbook includes hyperlocal influencers, SEO (Google My Business), and B2B2C partnerships (condos, real estate managers, beauty salons, churches).
  • 🚀 Smart ranking engine: ordered provider lists that blend rating, completed bookings, 5★ share, review recency, distance, acceptanceRate and averageResponseTime.
  • 🎯 Missions & rewards (gamified): event-driven missions (COUNT_EVENT, STREAK_DAYS, WITHIN_WINDOW) with claimable rewards (coupons or loyalty points) — including provider-focused goals like “10 jobs in a month”. Idempotent progress tracking & safe claiming. :contentReference[oaicite:1]{index=1}
  • 💵 PIX payments & instant withdrawals: backend-issued PIX charge (BR Code/QR) and provider withdrawals, surfaced in booking success and earnings flows.
  • 📈 Provider performance metrics: acceptanceRate & averageResponseTime exposed in profile/dashboard and used by ranking & discovery.

• Additional Functionalities:

  • ⚡ Dispute resolution system: built-in mediation flow between client and provider, reducing churn and ensuring trust.
  • 🗂️ Service categories & dynamic catalog: flexible setup for one-off services (cleaning, ironing, organization) and packages (weekly, monthly).
  • 🛠️ Admin dashboard: monitoring KPIs (NPS, churn, retention, CAC/LTV), fraud alerts, and financial health of the marketplace.
  • 📍 Geolocation & smart matching: PostGIS-powered search to match clients with nearby providers, optimizing time and travel cost.
  • 🔔 Real-time notifications: WebSocket-powered updates for bookings, cancellations, payments, and disputes.
  • 📈 Scalable architecture: NestJS microservices, Redis caching, BullMQ queues, Prisma ORM, and cloud-ready deployment with Docker/Kubernetes.
  • 🆘 Safety & panic alerts: endpoints for incident/panic flow, enabling rapid assistance and auditability.
  • 🧲 Referrals + Loyalty + Missions: tracked referrals that credit loyalty points and can trigger mission conversions after first booking.
  • 🎟️ Reviews analytics & AI replies: detailed rating breakdown and suggested responses to reviews to help providers win trust at scale.
  • 🛟 Support & SLA: typed Support module (tickets, messages, SLA logs) prepared for help center and agent workflows.

• A fun language learning platform that uses local AI models (Whisper & Mistral) to give you real-time feedback on your speech and have conversations that make sense.

  Built with a fast FastAPI backend and a responsive React Native frontend. It’s all about making learning feel like a game and keeping users engaged.

I have a strong base in how software should be made and follow good practices to make sure everything is high-quality and easy to maintain.

• AI/ML Lifecycle: I know how to handle the whole process, from getting data ready and figuring out what features are important, to designing, training, and deploying models.

• Clean Code & Design Patterns: You can see this in projects like `nestjs-dependency-inversion`, where I use smart ways of organizing code like Dependency Inversion.

• Agile Methodologies: I have experience with project plans and working in short cycles (like in the Bilingui-AI project), which helps in delivering things regularly and staying on track.

I'm really into open-source and always excited about new challenges. Feel free to check out my projects, suggest improvements, or reach out if you want to collaborate.