Hi, I'm Erik
Currently creating behavioral contract testing for AI-powered applications — custom Playwright matchers that validate non-deterministic AI output in CI against live endpoints. 20+ years spanning COBOL mainframes to AI-native test engineering.
118 Tests in CI
8 custom Playwright matchers validating AI behavior against live AWS Bedrock endpoints. Every PR. Every merge. No exceptions.
22,000+ Students
Built a K-12 LMS in 100 days serving 63 schools. FERPA-compliant, 155+ Row Level Security policies, 9.85/10 CTO assessment.
70% Automation
Plain English to executable Playwright tests. NLP-driven test generation with 95% parsing accuracy in under 5 seconds.
Systems I've built from scratch
From AI-native test infrastructure to education platforms serving 22,000+ students.
The Directive Platform
A three-component, spec-driven, AI-native test engineering system. CurioEVE (AI CLI, 60+ commands), SECUR-T (behavioral contract testing, 8 custom matchers), and M-O (test management platform). Built at a cybersecurity company, running in production CI against live AI endpoints.
Casiana
A full K-12 Learning Management System serving 63 schools and 22,000+ students. Built in ~100 days as a zero-revenue gift to a national education network. Next.js 16, React 19, Supabase, FERPA-compliant with Row Level Security on every table.
CasianaAI
AI-native test intelligence platform. NLP engine for natural language to test generation, intelligent model routing (60% cost reduction), self-healing tests (60% recovery rate), predictive failure prevention (90% accuracy). Named after my daughter.
The testing tools that don't exist yet
I find the gap, build what's missing, and ship it in production. 20+ years of creating what the industry is still debating at conferences.
Behavioral Contract Testing
8 custom Playwright matchers that validate non-deterministic AI outputs against behavioral contracts — not exact strings. Running in production CI against live AI endpoints.
680-Run Validation
Every flaky test fix runs 680 consecutive times before it's called done. Statistical confidence, not hope. Root cause classification and predictive hardening across the suite.
95% CI Speedup
15 lines of bash. Zero dependencies. Zero supply chain risk. Path-based CI gating that cut validation time from 8 minutes to 39 seconds on E2E-only PRs.
Spec-Driven Pipeline
Closed-loop from Jira ticket to passing test to coverage gap analysis. The Directive Platform automates the journey from requirement to validated code.
AI Trust Score
The Brain vs. Body framework — AI evaluation measures model quality, behavioral testing measures product behavior. Combined: the AI Trust Score. Most teams only measure half.
NLP Test Generation
Plain English requirements → executable Playwright tests. 70% automation rate, 95% NLP parsing accuracy, under 5 seconds. CasianaAI makes test engineering accessible.
Let's build something together.
Have a testing challenge? Let's talk.