Senior .NET Backend Engineer | AI Engineering | Microservices

Joan Antoni Serra Castillo

Computer Engineer with 8+ years of backend experience, specialized in AI engineering and enterprise distributed systems. I design end-to-end AI platforms, orchestrate multi-LLM agents, and deliver automation with measurable business outcomes through pragmatic architecture, clean code, and high-speed delivery.

  • 8+ years in backend and enterprise AI delivery
  • Risk automation impact: 8-9 hours/day saved
  • End-to-end AI platform: .NET 9, Next.js 15, Orleans
  • MCP, RAG, and distributed agent workflows in production
  • Remote / hybrid availability
  • Microsoft ISV Hackathon Madrid 2024 winner
Portrait of Joan Antoni Serra Castillo

Professional experience

End-to-end ownership across AI strategy, backend engineering, cloud migration, and enterprise reliability.

Jun 2025 - Present

AI Engineer | PAYRETAILERS (Remote, Barcelona, Catalonia, Spain)

  • Designed and developed an end-to-end enterprise AI ecosystem with AI-Core (.NET 9 REST API), AI-Front (Next.js 15/React 19), and AI-Agents-Semantic-Base (Orleans + Semantic Kernel).
  • Led Risk Agent automation flows across n8n, LLMs, and secure API integrations, achieving a daily reduction of 8-9 hours in manual operational work.
  • Engineered a scalable distributed runtime with Orleans 9.2 and Semantic Kernel, orchestrating 10+ LLM models and 40+ tools focused on fintech use cases.
  • Pioneered Model Context Protocol integration over SSE with custom MCP clients and servers, enabling secure dynamic tool discovery and enterprise system connectivity.
  • Implemented semantic retrieval workflows with PostgreSQL + pgvector (1536 dimensions) to improve context quality and response relevance in AI-assisted operations.
  • Delivered real-time AI interfaces with streaming chat, Markdown/KaTeX rendering, and an interactive workflow editor powered by XYFlow and Redux API slices.
  • Established an AI-driven SDLC workflow to convert functional requirements into production-ready code with faster delivery cycles and maintained engineering quality.
  • Implemented Databricks governance workflows and n8n orchestrations connecting Salesforce, Jira, and internal APIs for auditable automation at scale.
  • Co-led strategic backend migration to AKS and optimized Docker deployment patterns for highly available AI microservices.

Tech: .NET 9, ASP.NET Core, C#, Semantic Kernel, Orleans, Azure OpenAI, Anthropic, PostgreSQL/pgvector, Redis, n8n, Databricks, AKS, Docker, OpenTelemetry.

Mar 2023 - Jun 2025

Backend Software Engineer (.NET | AI & Microservices) | ZINANTIS / FASTPAYHOTELS

  • Designed and developed scalable microservices in .NET Core and .NET Framework for a high-traffic booking engine, applying Clean Code and SOLID principles.
  • Integrated Azure OpenAI and Semantic Kernel to automate extraction, transformation, and classification of unstructured data (Excel, PDF, and other sources).
  • Developed internal assistants and agent-based workflows using vector stores and semantic memory to support real-time operational decisions.
  • Built robust third-party integrations (REST and SOAP), transforming external data flows into stable internal domain models.
  • Optimized critical SQL Server queries, reducing response times and improving throughput in core business processes.
  • Collaborated closely with Product and DevOps teams under Agile/Scrum to ensure fast and reliable delivery cycles.
  • Provided technical leadership to junior developers while defining backend architecture and communication/data flow patterns.
  • Managed source control in Azure DevOps with GitFlow and reinforced quality through pull request reviews.

Tech: .NET Core, .NET Framework, Azure OpenAI, Semantic Kernel, Azure Service Bus, SQL Server, Azure DevOps, REST APIs, SOAP, Swagger.

Sep 2021 - Mar 2023

Full-Stack Senior Developer | IZERTIS (Remote)

  • Owned development and maintenance for an ecosystem of seven interconnected applications built on .NET Framework and .NET Core.
  • Improved SQL Server performance through indexing plans, query tuning, and stored procedure optimizations for high-volume business workflows.
  • Designed microservices patterns to improve modularity and reduce coupling across core business capabilities and shared infrastructure concerns.
  • Integrated asynchronous messaging with RabbitMQ to increase throughput and reliability in event-heavy processing pipelines.
  • Implemented and maintained background services for time-critical business processing, ensuring consistent execution and operational continuity.
  • Refactored and modernized legacy .NET codebases for better maintainability, compatibility, and production performance.
  • Redesigned WPF desktop interfaces to improve usability and reduce friction in internal user journeys.
  • Worked within Agile/Scrum methodologies with Product Managers and QA teams, using Jira to optimize delivery cycles and team coordination.
  • Provided mentorship to junior developers, reinforcing best practices in backend development and software design patterns.

Tech: .NET Framework (4.7.2/4.8), .NET Core, SQL Server, RabbitMQ, WPF, Windows Services, REST APIs, SOA, PostgreSQL, Jira.

Apr 2019 - Sep 2021

.NET Developer / Team Lead (last 6 months) | GRUPO DYNASOFT

  • Delivered backend solutions in .NET Core and .NET Framework focused on integrations and operational performance for business-critical systems.
  • Implemented REST and SOAP APIs connecting internal platforms with Dynamics NAV 2013 and external providers.
  • Optimized SQL Server and MySQL workloads with indexing and query tuning to support critical daily operations.
  • Worked as a full-stack developer integrating .NET backend services with React frontend interfaces for cohesive delivery.
  • Maintained and refactored legacy systems in PHP (CodeIgniter 3) and Windows Forms, preserving reliability while reducing technical debt.
  • Integrated Microsoft Graph API to automate communication and productivity workflows across internal teams.
  • Served as Team Lead in the final period, coordinating developers, conducting code reviews, and aligning technical execution with business requirements.
  • Collaborated closely with business and product teams to align technical implementation with functional requirements and operational priorities.

Tech: .NET Core, .NET Framework, SQL Server, MySQL, REST, SOAP, React, Razor, Windows Forms, WCF, Microsoft Graph.

Apr 2018 - Feb 2019

.NET Developer | QUONEXT / CEI EUROPE

  • Developed a WPF desktop application using the MVVM pattern, integrating SQL Server and WCF web services.
  • Built a REST API with n-tier architecture and created reusable NuGet packages for modularization.
  • Developed web applications using JavaScript (jQuery) inside an MVC architecture.
  • Implemented client-control add-ins for Microsoft Dynamics NAV, extending ERP platform capabilities.
  • Gained hands-on experience with Agile Scrum practices and collaborative multi-team delivery environments.

Tech: .NET Core 2.0, WPF, MVVM, SQL Server, WCF, REST APIs, NuGet, JavaScript (jQuery), MVC, Dynamics NAV.

Highlighted projects

Key systems and initiatives delivered across PAYRETAILERS and previous stages.

AI-Core (Enterprise Backend API)

Enterprise .NET 9 API with 200+ endpoints, 39 controllers, 48+ services, 53+ DTOs, and Liquibase-managed evolution. Covers chat streaming, agents, workflows, dashboards, and external enterprise integrations.

AI-Front (Next.js 15 + React 19)

Modern UI with real-time AI chat, streaming response rendering, XYFlow workflow editor, widget modes, and RTK Query architecture across 25+ API slices and 64+ component directories.

AI-Agents-Semantic-Base (Distributed Runtime)

Orleans + Semantic Kernel runtime with 69 grain interfaces, 44 production tools, multi-LLM support (GPT and Claude), MCP integration, Redis-backed state, and resilient orchestration.

Risk Agent Automation

Specialized risk operations automation using n8n, LLM orchestration, and secure system integrations (Salesforce, Jira, Databricks, internal APIs), delivering a daily reduction of 8-9 hours in manual workload.

Virtual POS System for Restaurants (Final Year Project)

Designed and developed a web-based Virtual POS using .NET Core and microservices, with REST API integration and real-time transaction/order management flows.

Platform architecture and scale

Consolidated metrics and architecture points from real enterprise delivery at PAYRETAILERS.

API layer

200+ endpoints, 39 controllers, 53+ DTOs, 58+ entities, and 15+ Liquibase migration versions.

Frontend layer

Next.js 15 + React 19, 64+ component directories, 25+ API slices, and node-based flow editing.

Agent layer

Orleans 9.2 runtime with 69 interfaces, 44 tools, YAML-driven config, and dynamic tool discovery.

LLM and MCP

Multi-LLM orchestration (10+ models), Semantic Kernel tool calling, and MCP client/server integration.

Data and governance

PostgreSQL/pgvector, SQL Server (legacy), Redis clustering, Databricks governance, and n8n automations.

Security and observability

JWT, Azure AD, OAuth2, Key Vault, OpenTelemetry, Azure Monitor, App Insights, and Polly resilience.

AI SDLC workflow

Our agentic development workflow uses HIP (Human Intervention Points) at critical control gates to preserve speed, quality, and production robustness.

1. Discovery and design

Product and architecture agents transform requirements into technical options, effort maps, and risk assumptions. HIP gate validates scope and architecture direction before build.

2. Agentic implementation

Coding agents generate modular code, tests, and migration scripts in parallel streams with shared constraints and coding standards across repositories.

3. HIP critical gate: security

Human review is mandatory for auth flows, secrets handling, external integrations, and data exposure boundaries before merge to protected branches.

4. HIP critical gate: contracts

API and data contract checkpoints verify backward compatibility, schema evolution strategy, and rollback safety for distributed services.

5. Validation and release

Automated quality agents execute unit, integration, and regression suites. HIP release gate validates readiness, error budgets, and observability coverage.

6. Production feedback loop

Monitoring agents summarize incidents and performance drifts into prioritized backlog updates, enabling fast and robust continuous delivery.

Technical stack

Capabilities aligned with the updated CV and production project experience.

AI & autonomous systems

MCP 0.4, agentic workflows, Semantic Kernel, Azure OpenAI, Claude, RAG, prompt engineering.

Backend & core

.NET 9, C#, ASP.NET Core, Orleans 9.2, Entity Framework Core, LINQ, multithreading.

Data & performance

PostgreSQL (pgvector/Citus), SQL Server (expert T-SQL, Dapper), Redis clustering, Databricks.

Architecture & cloud

Microservices, clean architecture, event-driven systems, Azure AKS, Key Vault, Blob, Docker.

Frontend & UI

Next.js 15, React 19, TypeScript 5, Redux Toolkit, Tailwind CSS, XYFlow, SSE streaming.

Automation & integrations

n8n, Jira, Salesforce, SharePoint, Microsoft Graph, ETL/ELT, OpenTelemetry-based monitoring.

Education & languages

Universitat de les Illes Balears

Bachelor's Degree in Computer Engineering (Sep 2013 - Jan 2018)

Focused on software engineering, architecture, and database optimization. Final project: virtual POS system for restaurants using .NET Core, microservices, and REST APIs.

Languages

  • Spanish: Native
  • Catalan: Native
  • English: Professional working proficiency

Contact me

If you want to get in touch, send an email directly and I will reply as soon as possible.

For collaborations or interviews, LinkedIn and email are both active.