Angel Campa

Four years at Vanco, from SDET to Principal. The numbers below are the ones another person can vouch for, because other people were in the room. They are internal figures about an employer's codebase, so the source for them is me.

• Frontend coverage from nothing to 93.34 per cent of lines across three Vue 3 and TypeScript applications, one of them roughly 450 single-file components and 280 TypeScript modules, with per-pull-request coverage workflows in CI.
• Backend coverage from zero to about 55.8 per cent of combined Go statements, across roughly 23,000 statements and 740 production files, with integration fakes, a coverage ledger and a pull-request gate. It was deliberately wrapped before the 80 per cent goal rather than padded to reach it.
• End-to-end testing from zero to 35 green, concurrency-safe Playwright tests, plus a multi-persona harness with trip-level parallelism and payment helpers that record without charging.
• Over 5,000 automated API regression tests for the faith vertical, and a migration off Java Selenium onto Playwright in TypeScript that removed the flakiness.
• Coverage turned into delivery: a 23-bug fix bundle found through characterisation testing, and Go defects found through integration coverage and closed with regression tests.
• An agentic QA platform for two products, with manual black-box testing and Playwright authoring as separate skills, a shared governance pack, a multi-agent learning loop, and an expected-behaviour oracle built from the help centre.
• I also ran the company-wide QA guild, reviewed other engineers' test automation across teams, and pushed the adoption of AI-assisted development internally.

A master's in computer science, and before it a law degree and an economics degree, and a year teaching maths, history, accounting and law at high-school and college level. People tend to make more of that than I do. My own guess is that it overlaps with the engineering by about ten per cent, and I am not the most impartial judge of it.

The graduate work was applied machine learning on financial time series: an LSTM for price prediction, a genetic algorithm tuning technical indicators, and a scraper that pulled records off a government site into MySQL.