I am Arthur Cabral,
a Software Engineer
based in Brazil.

At Amigo Tech, I began as a Junior Software Engineer in November 2021, was promoted to Software Engineer in June 2022, and to Senior Software Engineer + Team Leader in July 2024.

• Amigo Chat: I collaborated on building the real-time messaging system using Socket.io, Node.js, and AngularJS—contributing to socket architecture, Redis-backed message persistence, and Casl-based authorization modules. This solution cut internal team response times by 30%.

• Core Refactors: As a mid-level engineer, I led refactors of the scheduling, medical-record, and financial microservices—rewriting critical endpoints in TypeScript and optimizing PostgreSQL and Redis queries. These efforts boosted latency performance by up to 40% and improved code maintainability.

• Amigo Pay: From July 2024 onward, I joined a cross-functional squad to plan and implement Amigo Pay. I defined account onboarding/management workflows, integrated card terminals and PIX via webhooks, and reinforced security (OTP-based MFA, enhanced email validation, dynamic rate-limiting, RBAC). The module processed R$ 7.5 million in TPV in its first year and reduced vulnerabilities in the core services.

• Platform Development: Built a unified marketing-automation platform using React, Next.js and TypeScript on the front end, with Node.js/Express and FastAPI (Python) back end, PostgreSQL and Redis for data storage, and Heroku for deployment.

• API Integrations: Implemented OAuth2 integrations with Meta, Google Ads, and Pinterest APIs to automate campaign, ad-set, creative and ad creation. Normalized performance metrics (CTR, CPC, conversions) into relational schemas and cached them in Redis.

• AI Optimization: Developed a GPT-4 recommendation engine that dynamically suggests creative improvements and audience optimizations, resulting in a 20% increase in ROAS and an 80% reduction in manual setup time during pilot tests, with five accounts onboarded in the first month.

At Rekon, I developed a face-recognition and image-processing platform using Node.js/Express and React, with PostgreSQL for persistent storage, Redis for caching, and Docker for containerization. I engineered an end-to-end image upload pipeline (resizing, normalization), integrated Face++ API and AWS Rekognition for accurate face detection and identity verification, and implemented a "selfie search" feature that returns matching results in under 5 seconds.

At Retailsparks.AI, I designed and implemented a high-throughput data pipeline using Python (FastAPI) and AWS Lambda to ingest over 10,000 Reddit posts per day via PRAW, then applied Hugging Face's BERT model for real-time sentiment analysis. I persisted the processed data in PostgreSQL, built a React dashboard for interactive trend and buzz visualization, and orchestrated the entire ETL workflow to run reliably at scale. The MVP was validated by three retail clients—who leveraged the insights to boost community engagement by 25%—and the system maintained 100% data integrity with zero downtime during peak loads.

At UFRPE, I studied Computer Science and Systems Information, with a focus on Software Engineering and Enterpreneurship. I also completed a Master's in Computer Science at the same institution, where I specialized in Software Engineering and Web Development.