Hi there, I'm Paulo Silas de Campos!

I am a highly motivated Senior Software Architect | AI/ML & Infrastructure Engineer focused on building scalable and intelligent applications that really work. I enjoy tackling complex challenges and turning them into smart solutions, using modern development and cutting-edge machine learning.

Shipping end-to-end products from concept to scalable production with real users

Core Engineering & Production Systems

Swappy Financial - High-Throughput Web3 Architecture

Corporate On/Off-ramp platform (Fiat <> Crypto) processing real-time transactions.

• Distributed Architecture: Built a Web3 engine based on Microservices with asynchronous communication to ensure eventual consistency in on-chain operations.
• Hierarchical Deterministic (HD) Wallets: Implemented key management (TRC20) via address derivation, ensuring fund isolation and hierarchical idempotency.
• Idempotency & Security: Double-spending and replay attack protection using distributed nonce systems, ts-validation, and HMAC-signed webhooks.
• Infrastructure: Automated CI/CD pipelines using Terraform on Azure, ensuring immutable environments and Infrastructure as Code (IaC).
• Real-time UX: Three.js integration for real-time visualization of network fees and gas instructions (TRX).

LimpeJá Marketplace - Scalable Platform Engineering

Uber-like platform for cleaning services | 191 providers | 251 active clients | R$ 82,850+ GMV

• Real-time Matching Engine: Intelligent recommendation algorithm processing geolocation and availability in milliseconds for ultra-fast booking flows.
• FinOps & Payout System: Payment architecture integrated with PIX APIs, focused on financial transparency and instant withdrawals (<24h).
• Identity & Compliance (KYC): Image processing pipeline with OCR + Facial Recognition integrated into onboarding, ensuring LGPD (Data Privacy) compliance.
• Observability: Implementation of structured logging and metrics monitoring (SLIs/SLOs) to track acceptance rates and response times.

Bilingui-AI - Edge AI & Inference Optimization

Language learning platform with offline AI models

• Edge Computing: Migrated cloud models to Local Inference (Whisper/Mistral), reducing API costs by 100% and inference latency by 75%.
• High-Performance Backend: API developed with FastAPI optimized for audio streaming and parallel processing.
• Clean Architecture: Decoupled AI models from business logic, allowing seamless LLM swapping and maintainability.
• UX: Gamified learning experience with 100% offline real-time feedback processing.

Technical Stack & Expertise

• Infrastructure & DevOps: Terraform (IaC), Azure (AKS), Docker, Kubernetes (K8S), Helm, GitHub Actions, Nginx, Linux (Arch), Bash.
• Architecture: Microservices, Event-Driven, Hexagonal Architecture, REST/gRPC, Redis (Caching/Queue), RabbitMQ, Kafka.
• Languages & Frameworks: Python (FastAPI/PyTorch/TensorFlow), TypeScript (NestJS/React/Next.js), Go, Rust, C++, Node.js, Spring.
• Database: PostgreSQL, MongoDB, Redis, Pinecone (Vector DB), SQLite, DynamoDB, Cassandra, Elasticsearch.

Architectural Philosophy

I don't just write code, I design systems that scale. My priority is observability, resilience, and FinOps cost-effectiveness, ensuring that the infrastructure supports business growth without technical bottlenecks.

🤝 Let's Connect

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.

Open to: Senior Engineering Roles • Technical Leadership • Startup Advisory