Scalable Systems · Microservices · Cloud Native · AI Integration
Backend Engineer with over 6 years of experience designing scalable microservices and high-performance APIs in Python/FastAPI. I combine the rigor of Hexagonal Architecture / Clean Architecture with the agility of Cloud Native services, and I build integrations of AI agents at the SDK level (MCP, Anthropic & Gemini).
📍 Lima, Peru — Open to Remote (Global)
🎯 Job Match — Python · FastAPI · Gemini · pgvector · Airflow · React An end-to-end AI-powered job matching pipeline. It collects job postings from legal sources every 12 hours, extracts requirements using an LLM (Gemini) into a validated Pydantic schema, calculates embeddings, and performs semantic scoring (using LLM) against a professional profile, exposing matches with strengths and risks via API. Built on Clean Architecture (domain → application → infrastructure → interfaces), with PostgreSQL + pgvector, JWT authentication, Airflow orchestration (DAG every 12 hours), Docker + Alembic, Prometheus metrics, and a React frontend (Vite + TS).
🤖 Starbucks AI Agent — TypeScript · Claude & Gemini SDKs · MCP · ChromaDB A production AI backend with agentic workflows and RAG integration built at the SDK level (without LangChain). It exposes MCP servers (FastMCP) for agent-tool interoperability, a hexagonal/DDD architecture, approximately 80% test coverage, and Prometheus/Grafana/Loki observability.
⚡ Realtime Alert System — Go · Kubernetes · Terraform Event-driven alert service in Go with Clean Architecture, semantic versioning, CI/CD with GitHub Actions, and Infrastructure as Code (Terraform) for AWS.
🔐 OAuth2 Microservices Library — Python · Keycloak Service-to-service authentication library (OAuth2 / JWT) with Keycloak for microservices architectures.
⛓️ Mini Blockchain — Python Exploring immutable data structures and distributed cryptography.
"I build for scale, maintainability, and resilience."
-
Modernization: Critical migrations from .NET monoliths to FastAPI microservices using the Strangler Fig Pattern, without disrupting production.
-
AI Integration: Design of intelligent agents and RAGs (ChromaDB) at the SDK level, with MCP servers (FastMCP) for tool interoperability.
-
Security First: OAuth2/JWT libraries and secret management with HashiCorp Vault.
-
Design Patterns: CQRS, Event-Driven, DDD, and SOLID as the foundation for maintainable systems.
| Project | Tech Stack | Key Outcome |
|---|---|---|
| Real-time Alert Modernization | Python, FastAPI, GCP | Latencia de API reducida ~95% (60s → <3s) vía procesamiento event-driven. |
| Panic Alert Integration | C# .NET, Webhooks | Integración centralizada para Genetec Security Center 5.13. |
| Event-Driven Core | Kafka, RabbitMQ | Comunicación asíncrona confiable en sistemas de alto tráfico. |
| Automated Pipeline | GitHub Actions, Terraform | IaC para despliegues Zero-Downtime. |
- Languages: Python (FastAPI, Django, Flask), TypeScript (NestJS, Node.js), Go, Java, C# / .NET
- AI / MCP: MCP servers (FastMCP), Anthropic & Gemini SDKs, RAG, embeddings, pgvector, ChromaDB
- Cloud & DevOps: AWS (Lambda, S3, API Gateway, RDS), GCP, Docker, Kubernetes, Terraform, GitHub Actions
- Data: PostgreSQL, MongoDB, Redis, pgvector, Pandas
- Messaging: Kafka, RabbitMQ, Celery
- Architecture & Practices: Clean / Hexagonal Architecture, CQRS, Event-Driven, DDD, SOLID, Serverless, TDD
- Security: OAuth2, JWT, HashiCorp Vault
- 📧 [email protected]
- 🔗 LinkedIn · GitHub
- 🏎️ Passionate about Sim Racing (Moza Racing setup)