Monday, October 20 marked the third edition of Urudata Software’s 2025 Tech Talks, a space where ideas and practice meet to explore how artificial intelligence and automation are redefining software development.
This edition featured Julia González, Richard Marino, Matías Saravia, and Agustín Rosado, who shared their experiences and demonstrations of tools that boost productivity, enhance collaboration, and connect development processes with new AI capabilities.
From Design to Production: Builder and Cursor
Julia González (Junior Designer) and Richard Marino (Software Analyst) kicked off the session presenting the tools Builder and Cursor and their potential uses to accelerate software development from initial design to functional code.
Builder is a visual web development platform that integrates directly with Figma. It allows designing interfaces visually and automatically generating code—combining design and development in a single platform—while supporting multiple frameworks, real-time collaboration, and integration with external tools.
Cursor, on the other hand, brings AI collaboration directly into the development environment. It allows interacting with the repository, understanding project patterns, and generating code consistent with the existing style. It also includes a Background Agent, which can execute tasks in the background—like creating tests, documenting, or reviewing PRs—while the developer continues working.
Both speakers highlighted the potential of AI-first workflows, where well-designed prompts and iterative tool usage enable moving from prototype to production with greater speed and consistency.
MCP Servers: the bridge between LLMs and the real world
In the second talk, Matías Saravia (Software Analyst) introduced the Model Context Protocol (MCP), a new open standard aimed at connecting language models with external data, tools, and systems.
The protocol defines a symmetric client-server architecture that allows any AI application to communicate with an MCP server, dynamically discovering its capabilities (tools, resources, prompts, utilities). This eliminates the need for manual, proprietary integrations, reducing duplicated effort and security risks.
The speaker demonstrated practical examples of tools that can perform actions such as creating files, reading repositories, or querying databases, showing how interoperability between models and real infrastructures will be key in the next generation of intelligent applications.
Process Automation with GitHub Actions
The session concluded with Agustín Rosado (Junior IT Support), who reviewed how GitHub Actions enables the automation of complete CI/CD flows, from building to deploying applications, integrating testing, quality analysis, and AI-assisted reviews.
He also highlighted its flexibility for different environments—cloud, on-premise, or hybrid—and its integration with the GitHub ecosystem, where workflows can be configured in YAML files and combine multiple jobs in parallel.
With live examples and demos, he showcased concrete use cases: automated builds and tests, AI-assisted code reviews, and automatic documentation generation.
Continuous Learning
With this new edition, the Tech Talks continue to serve as a meeting point between Urudata Software teams and the emerging technologies driving innovation, enhancing daily tasks.
This third edition reflected a clear trend: automation and AI are no longer isolated tools, but integral parts of the daily workflow—a premise that, as an AI-First company, we adopt and support in every project and solution.
