Bảo Tô

At GTV, I tested a multi-country digital banking and payments platform across web, mobile, and backend APIs. I built Playwright and Cypress automation in GitLab CI, reducing regression cycles from roughly eight hours to five.

I've combined seven years of QA work across banking, fintech, logistics, telecom, and real estate with practical AI/ML evaluation research. My work includes API and SQL validation, root-cause analysis, test strategy, defect management, and automation frameworks built from scratch.

Through TraceCase, I designed and shipped a React SaaS application for anti-hallucination test-case generation. It uses a maker-verifier approach to distinguish evidence-backed EXPLICIT outputs from justified INFERRED assumptions and calculates a grounding ratio for output reliability.

I'm building an LLM evaluation portfolio with DeepEval and exploring cross-model benchmarking through OpenRouter, along with Claude Code subagents for debugging, test strategy, architecture, and code review.