Software testing for the agentic era.
Hi, I’m Nick Baynham.
I’m a software testing and automation engineer focused on the next generation of quality systems: AI-assisted test design, agentic workflows, API-first automation, and practical quality signals that help teams ship with confidence.
Featured projects
All projects →Universal Testing Language
A test-intent DSL that translates to multiple automation frameworks. One specification, many runtimes.
Agentic Testing Workflow Prototype
An AI agent that helps with test generation, maintenance, and triage - with explicit guardrails and a human in the loop.
API Automation Framework
A pragmatic Python framework for API contract testing, fixture-driven data setup, and CI-friendly reporting.
Latest writing
All posts →The bugs Test Commander is catching in Marketing Commander: a field report from a greenfield build
Test Commander has been attached to a greenfield product since before its first commit. This is a taxonomy of the real bugs it has surfaced - substring traps, lost-update races, priority-inverted requirements, a ninety-minute CI hang - and the three near-misses where a bug slipped past the developer and the passing test suite and the CI, and only an independent review caught it.
July 22, 2026 · 16 min read
Generating a UI test suite for a live app: eight failures that were really findings
Asked to generate a UI test suite for a live Angular app, the honest answer was that the auto-generated scaffolds were a non-runnable skeleton. The real suite had to be hand-authored and debugged into green against the running system. It went from 30 to 41 passing tests across eight rounds of failure, and every failure was the application teaching the test something true about itself: a blank edit form, unlabelled number inputs, an anchored URL validator, a list that does not render reliably, and a backend that mishandles concurrent writes.
June 13, 2026 · 14 min read
Test Commander against a live API: from a Swagger URL to a failing quality gate
A user dropped a single Swagger URL and asked for testing. In one session, Test Commander drove the full pipeline against a live FastAPI app: ingest five knowledge sources, review requirements, run three live exploration charters through Playwright MCP, generate BDD, ship a runnable regression suite, and produce an evidence-backed quality report. The result was not a green dashboard. It was a FAIL gate backed by six reproducible defects, two of them critical security bugs, and one clean root cause underneath all of them.
June 13, 2026 · 16 min read
Skills snapshot
Test automation architecture, API testing, Playwright, Python and TypeScript SDETs, CI/CD, AI-assisted test design, and pragmatic adoption of agentic workflows.