I am Antonio Rodriguez Farias, a QA Automation Engineer. I test AI systems and the software around them: LLM agents, APIs, web applications and data pipelines. I build the automation that tells a team, on every commit, whether the product still does what it should and stays safe while doing it.
LLM evaluation and AI agent testing
Teams are shipping LLM agents faster than they can test them. I design automated evaluation harnesses that measure quality, safety and security on every change: golden datasets, DeepEval metrics, Promptfoo regression checks and Garak red teaming, wired into GitHub Actions.
- Tool use, grounding and forbidden-content checks that are deterministic and reproducible
- Prompt injection, jailbreak, system prompt leakage and access control testing
- A detection gate that fails the build if the tests stop catching a known-vulnerable agent
Test automation across the stack
The same engineering discipline applies to every layer of a product.
Backend
API circuits in Python with authentication, database verification and automatic cleanup.
Frontend
Selenium, Playwright and Behave BDD across Chrome, Firefox and Edge.
Data
ETL validation, SQL reconciliation and NoSQL assertions.
Performance
Latency budgets in every suite and percentile-based load tests.
Pipelines and environments
Tests only protect a product when they run. I build unattended pipelines with GitHub Actions and Jenkins, and I cover cross-platform support with disposable machines in Vagrant and VirtualBox.
Stack
Python, pytest, DeepEval, Promptfoo, Garak, LangChain, LangGraph, FastAPI, Selenium, Playwright, Behave, SQL, Firestore, Pentaho, GitHub Actions, Jenkins, Vagrant, Linux, macOS and Windows. MSc in Big Data, Postgraduate Diploma in Defensive Cybersecurity and Diploma in Machine Learning.
Read more about me, or get in touch at antonio@rodriguezfarias.com, on LinkedIn or on GitHub.
