Background
From Developer to Architect My journey began in software development, building web portals and management systems, which gave me the "white-box" technical foundation that defines my approach to quality today. Unlike traditional testers, I transitioned into high-complexity automation early on, spending over a decade designing frameworks from scratch for enterprise data virtualization and cloud architectures at companies like GlobalLogic and OpenGov. Bridging Dev and Ops I have always operated at the intersection of development and operations. I spent six years leading validation strategies for public infrastructure software, managing everything from load testing georeferenced data to implementing CI/CD pipelines across multi-cloud environments. Simultaneously, I maintained my full-stack engineering skills through senior roles in FinTech and eCommerce, ensuring I never lost touch with the code I was testing. The Academic Pivot Currently, I am leveraging this experience to bridge the gap between industry bottlenecks and academic innovation. While working as a Senior AQA, I am pursuing my Ph.D. to develop an AI Agent based on the Model Context Protocol (MCP). My goal is to move beyond fragile "record-and-playback" tools and build autonomous systems capable of interpreting user stories and generating robust test code independently.
Why I Started
Solving the Agile Bottleneck I founded this initiative because Quality Assurance remains the critical bottleneck in modern software development. After 15 years, I saw that traditional automation tools are too fragile and manual to keep pace with the volatility of daily sprints. The Autonomous Solution I am building the solution I always needed: an autonomous AI Agent that simulates a human SDET. By leveraging GenAI and the Model Context Protocol (MCP), my system interprets user stories to generate and maintain resilient test code automatically, replacing manual scripting with intelligent orchestration.
What I'm Working On
Architecting Intelligent Quality I am currently a Senior AQA at GlobalLogic, where I design high-performance automation frameworks for global marketing platforms using Playwright and Python. My role focuses on integrating quality gates into CI/CD pipelines to support complex microservices architectures. Building the Autonomous SDET Simultaneously, I am advancing my Ph.D. research to revolutionize how testing is performed in agile environments. I am developing an AI Agent based on the Model Context Protocol (MCP) that goes beyond code generation; it autonomously interprets ambiguous user stories, plans testing strategies, and executes self-repairing test code, effectively simulating the cognitive role of a human SDET.
Allprocar
A comprehensive automotive management platform connecting vehicle owners with workshops and mechanics to streamline repairs, maintenance tracking, and service scheduling.
Automotive Services / Workshop Management Platform
Visit WebsiteBiggest Struggle
Implementing robust guardrails to constrain the AI's context solely to Allprocar's business logic, effectively filtering out out-of-domain queries like general knowledge.
Trying to Learn
I am focused on mastering the development of a domain-specific AI Agent for Allprocar. My primary goal is to learn how to engineer robust guardrails and context management systems that keep the agent strictly focused on automotive services (appointments, repairs, vehicle history) while preventing it from answering unrelated queries. I also want to understand how to seamlessly integrate these AI capabilities with my Go/GraphQL backend to query database records dynamically.
Help Needed
I need technical guidance on designing strict guardrails and context management systems for the Allprocar AI Agent to ensure it stays within the automotive domain. Investors also: I am seeking introductions to early-stage investors (Angels or pre-seed VCs) with a focus on Vertical SaaS or the Automotive Industry to help accelerate our development and go-to-market strategy.
Can Offer
I can offer insights on engineering domain-specific AI Agents, sharing strategies for implementing strict guardrails to keep LLMs focused on business logic (avoiding hallucinations). I can also collaborate on the technical architecture for integrating AI with Go/GraphQL backends, specifically discussing patterns for real-time data querying and context management in conversational interfaces.
Read Matías's full interview
Discover their complete journey, challenges, and insights in our in-depth interview.
Read Interview