Jose Manuel Casal

End-to-end design and development of a production backend system integrating LLM-based analysis into real-world domain workflows.

Not a prototype. Not a proof of concept. It's a complete system with Clean Architecture (.NET 8), Dockerized pipeline, and a configurable analysis engine with no hardcoded instructions.

AI behavior —evaluation criteria, domain rules, decision weights— is defined in configuration, not in code. Any change in business logic is applied without touching the system. Every model decision is auditable.

The architecture: Core / Application / Infrastructure / Worker. A Chrome Extension for data capture. An SPA in Alpine.js. A processing pipeline running with Quartz on Docker.

This is what integrating AI into a real system means: not just pasting an API call into a controller, but designing from the start for the AI component to be interchangeable, configurable, and auditable like any other system module.

Stack: .NET 8 · Claude API · SQL Server · Docker · Quartz · TypeScript · Alpine.js · Clean Architecture

·· AI-Integrated Backend System ··

AI-integrated .NET 8 backend system. Clean Architecture — Core / Application / Infrastructure / Worker. Settings-driven LLM behavior, zero hardcoded prompts. Configurable scoring model with domain rules defined at runtime. Chrome Extension data capture. Alpine.js SPA. Dockerized pipeline with Quartz scheduler. Full auditability on every AI decision. Production-grade AI integration. LLM · RAG · Agentic workflows

Backend consultant for 4 years on a multitenant SaaS platform for regulated financial advisory (MiFID2). Real production environment, with regulatory consequences when something fails.

The work was not a one-off project: it was sustained involvement across multiple modules — account and financial product management, portfolio reporting, advisory engine, digital signature, document management.

Most of the work: refactoring critical components under production constraints, performance optimization, and improving per-tenant observability. In a regulated environment, you can't just make changes and see what happens. Every decision must be justified, traced, and validated.

4 years on the same system give you something that cannot be faked: knowledge of every design decision, every technical debt, every fragile point. And knowing when to touch and when not to touch.

Stack: .NET 6/8 · ASP.NET Core · SQL Server · MongoDB · Docker · RabbitMQ · Redis

·· Backend Consultant — Wealth Management Platform · Fintech ··

Backend consultant on multitenant SaaS platform for regulated financial advisory (MiFID2). 4 years, multiple modules, production environment. Account and financial product management, portfolio reporting, advice engine, digital signature, document management. Refactoring critical components under production constraints. Performance optimization and per-tenant observability. Distributed team, strict compliance requirements. Fintech backend, wealth management platform. .NET 6/8 · ASP.NET Core · SQL Server · MongoDB · Docker · RabbitMQ · Redis

Sole technical lead for the comprehensive management system of an air handling operator. Designed, built, and maintained in production for over 4 years.

The system is the complete operational backbone of the business: flight and operations management, manifests, tariff engine, automated billing, warehousing, and accounting export. No business module is untouched by this system.

No team. No handoff. A single engineer who knows every layer — from the domain model to Azure deployment — and keeps the system running under real operational pressure.

This has a concrete implication: when something fails in production before a flight, there's no one to escalate to. That forces building well from the start and knowing the system in depth. 4 years of total ownership — every design decision, every technical debt taken on or avoided, every Azure architecture decision: made with full knowledge that you'll have to live with it.

Stack: .NET 5|6|8 · ASP.NET MVC · REST API · Azure App Service · Azure SQL · xUnit

·· Backend Engineer — Integral Air Handling System · Logistics ··

Sole technical owner. Full-cycle management system for air handling operator. Flight and operations management, manifests, tariff engine, automated billing, warehousing, accounting export. 4+ years in production. Single engineer ownership — design to Azure deployment. No team, no handoff. Ground handling software, airline operations, cargo management. .NET 5|6|8 · ASP.NET MVC · REST API · Azure App Service · Azure SQL · xUnit