Jimmy Auris Castillejos

• Led the consolidation of the internal platform by designing and implementing a complete flow with APIs (FastAPI) and frontend (Next.js), centralizing data capture and reporting in dashboards. • Implemented CI/CD pipelines with GitHub Actions and containers (Docker, Kubernetes), achieving a ~80% reduction in deployment time for new features. • Designed and applied FinOps practices during migration and operation on AWS (EC2, Lambda, RDS), optimizing cloud operational costs by 25%. • Built serverless processes (AWS Lambda + EventBridge) for massive calculations that fed Power BI dashboards, reducing processing latencies by up to 40% in critical workflows. • Managed database migration from MySQL to PostgreSQL (AWS DMS) and administered the RDS instance, improving platform resilience and analytical capacity.

• Transformed ad-hoc scripts into a robust system, migrating SPs to ORM (SQLAlchemy) and refactoring to hexagonal architecture, reducing critical query load by up to 70%. • Developed an internal system for investor management that supported an increase in managed capital of ~S/100 million, accelerating reports and real-time decision making. • Implemented Python automations (pandas, NumPy) for tax reporting and reconciliations, saving ~15 hours/week on repetitive tasks and significantly reducing manual errors. • Technically led the CRM redesign and integration of conversational bots (n8n), resulting in a 15% increase in commercial disbursements (~S/2 million additional). • Migrated and self-managed internal tools (n8n, NocoDB) with CI/CD pipelines, positioning the company at the forefront of commercial process automation.

• Digitized and automated tax reports and operational processes with Python, improving operational efficiency by ~60% and freeing up analysis time. • Implemented alerts and reports with pandas/Plotly integrated into internal dashboards, reducing process error incidents by ~20%. • Showed initiative to create scalable solutions, laying the foundation for migrations and systems that would later be formalized.

• Cleaning and structuring large volumes of unstructured data for coercive collection analysis, improving report accuracy by ~30%.