I'm currently employed as Senior Developer at Engineering Ingegneria Informatica (Milan, Italy) since November 2024.

Engineering is a Digital Transformation Company, leader in Italy and constantly expanding worldwide: 15.000 employees and more than 70 companies in 21 countries. I am working in the area of financial services, in particular maintaining a software product and migrating it in a microservices-AWS cloud architecture for an important client in the petroleum industry.

Cloud modernization of a complex management system

To date, my main professional activity has been contributing as a Senior Developer to the migration of a legacy management system commissioned by a major client in the petroleum industry and used across Italy, Spain, the Netherlands, and Belgium. The project involved transitioning the system from a monolithic MVC architecture to a microservices-based architecture hosted on AWS.

Technology Stack: transition from Java-only, on-premises environment to modern stack based on Java, Angular, Docker and AWS.

Architecture and Infrastructure:

• replacement of the original monolithic MV architecture with a distributed cloud infrastructure based on REST microservices, batch processes, and serverless functions.
• Replacement of SOAP service-to-service communication with event-driven messaging, data streams, and REST-based web services.
• Migration of JSP and JavaScript-based front end to single-page web application.
• Adoption of AWS as the cloud provider,
• exposition of back-end functionality through REST APIs
• improved communication with external services through modern data-transmission protocols.

Software Quality, Defect Resolution, and Refactoring

Identification and analysis of existing flaws: implementation of fixes and refactoring activities to improve reliability, maintainability, and security of the existing business logic.

Automated Testing

• Introduction of automated testing from an initial coverage of 0%, to an approximate average coverage of 20%, with daily continuous growth.
• Coverage for critical business-service classes ranged 50%-80%, within the functionalities under my direct management¹.
• Since the migrated microservices generally involved limited sequential processing elaborations, API integration tests were mostly developed, leveraging in-memory H2 databases.
• Test scenarios were designed sticking to the following covering approach:
  • happy paths,
  • counterfactual cases,
  • controlling for different business conditions,
  • affected data-segregation verification.

AI-Assisted Development

• Introduction of AI-assisted coding practices into the development workflow.
• Developed custom multi-agent workflows to support repetitive, multi-step tasks.
• Integrated documentation and instructions for RAG-based solutions.

Local Development and Testing

• Replacement of the manual MVC setup based on Eclipse and the Maven plugin with Linux-based development environment running through WSL.
• Integration of LocalStack to emulate AWS services locally;
• adoption of AWS SAM to support serverless development and testing. I also introduced additional integrations for debugging, testing, and operating locally with cloud resources, helping streamline development and testing workflows.

DORA scoring

Passed from 5.6 Overall Performance score (against a 7.0 \(\pm\) 1.0 average in 2025 Industry baseline in Financial Services) to a 7.6 Overall Performance².

Documentation

Improvement of the original scarce and fragmented documentation by creating UML diagrams, functional documentation, and Markdown documentation for critical processes and workflows.

CI/CD Automation

Replacement of the manual delivery practices with a fully automated CI/CD pipeline based on AWS CodePipeline, CodeBuild, CodeArtifact, and CloudFormation.

Development Methodologies and Collaboration

• Scrum practices during the most active migration phases, where there was the need to complete intensive activities within the deadlines;
• Agile principles throughout the subsequent maintenance and improvement lifecycle.
• Both autonomous and peered programming; minimal code-review.

Limitations

• Extending test coverage beyond a certain level was not feasible for some functionalities due to their specific characteristics and constraints.
• Minimal code review process.

AI champion

I served as a reference "AI champion" within the business unit, providing mentoring and promoting the adoption and effective use of generative AI tools to improve productivity.

In this role, I also acted as a tester and analytical evaluator of the company’s custom tools based on RAG, multi-agent AI, and automation, characterized by segregation, tenancy, and a high level of customization. I conducted comparative analyses against alternative solutions as baselines such as custom RAG implementations and cloud services like Bedrock, while coordinating a survey-like assessment by collecting primary data from collaborators, documentation, and direct testing.

Research activities

I led and reported a PoC to evaluate the efficiency gains of purpose-built agents developed specifically to support the migration of a legacy application (based on Oracle stored procedures and Java Structs MVC architecture) to the new distributed architecture. The study defined structured evaluation methodology, assumptions, scope, and limitations, and estimated an approximately 40% increase in time productivity and a 10% reduction in the cost of the resources involved compared with the no-AI baseline (previous quote). The study was determinant in winning the engagement by demonstrating the required cost and efficiency gains against a baseline aligned with competitors’ estimates.

I also developed a demo RAG solution for internal documentation, which is detailed here.

Other activities and training

Collaborated with international colleagues, requiring effective English communication in both written and spoken form. Contributed to on-call support duties, including nights and weekends approximately once every 1.5 months. Participated in company-wide activities and training initiatives, while continuously engaging in both company-provided and self-directed learning. During this period, I successfully balanced my professional responsibilities with ongoing studies toward a Master of Science in Applied Artificial Intelligence, and also earned the AWS Solutions Architect certification.

Years: 2024-

1. Coverage was estimated using JaCoCo for the managed microservices only; managed serverless functions are excluded, although they are covered by automated tests (integration of such metrics will follow). ↩
2. https://dora.dev/quickcheck/?v=2025 ↩