• Engineered and deployed 6+ production-grade ML pipelines for a financial fraud detection platform using Python, TensorFlow and Docker, securing 99.2% uptime and optimizing model inference latency by 45% (reducing API response time from 250ms to 137ms). • Architected and developed “FinAdvisor AI,” a full-stack personal finance application built with Flutter with Google Cloud Platform backend, leveraging Open Finance data via Pluggy to deliver secure, AI-generated behavioral insights. • Developed an automated financial reconciliation system for Cimtal Construções utilizing Python and FastAPI, with n8n automations synchronizing live bank API data across 15+ accounts and eliminating 15 hours of manual labor weekly. • Architected a complex Python data pipeline for Instituto Dragão do Mar to aggregate, clean, and deduplicate over 500,000 student records from 400+ unstandardized databases into a unified BI source of truth in a PostgreSQL database. Skills: Python, FastAPI, Flutter, PostgreSQL, TensorFlow, GCP , Docker, Firebase, PostgreSQL.

• Designed the architectural redesign of the VIVO Monitoring Systems web client using React and TypeScript, integrating AI-driven analytics to elevate B2B client satisfaction from 80% to 98%. • Engineered a high-performance "Persistent-mode" for CCTV dashboards that dynamically unmounts non-essential UI components using React and TypeScript, reducing browser memory consumption by 60% during extended multi-camera monitoring sessions. • Spearheaded UX/UI design and front-end architecture of LAPISCO Academy using Figma, TypeScript and React, scaling the platform for 5,000+ concurrent users, reducing the production bundle size by 33.2%, and accelerating page load times by 48.4%. • Trained and deployed a YOLOv8 Computer Vision model for real-time pedestrian and vehicle detection, resolving 100% of tracking bugs and boosting processing speed from 18.08 to 24.34 FPS. Skills: React, TypeScript, Python, YOLOv8, OpenCV, Figma, Computer Vision, Local AI Models, LLMs, Keras, TensorFlow.

• Architected and deployed enterprise-grade, AI-driven agentic automation pipelines for B2B clients utilizing n8n, Claude Code, and Google ADK, reducing manual operational overhead by an average of 75% per client. • Engineered a custom n8n automation workflow integrated with the Meta Business API for WhatsApp, securely processing over 10,000 automated, AI-driven client interactions monthly with zero downtime. Used JavaScript code for custom JSON parsing. • Developed a centralized, high-performance web dashboard using React and TypeScript for clients to monitor AI agent performance in real-time, reducing support inquiries by over 40%. • Spearheaded the research and development of a "Super AI agent" prototype leveraging Paperclip and Opencode to autonomously handle complex marketplace integrations and workflow scheduling. Skills: React, TypeScript, n8n, Python, OpenClaw, Claude Code, Meta Business API, Paperclip, Opencode.