Test automation & infrastructure
AI-agent-driven end-to-end test suite
WhatDesigned, consolidated, and maintains the automated end-to-end regression suite for a consumer mobile app on Android and iOS, executed by AI test agents across real devices, emulators, and simulators
Scale~250 test jobs per day across two platform configurations; hundreds of distinct test cases
HowDeclarative agentic test cases, Python orchestration, shared vs. platform-specific case hierarchy, CI batch scheduling, automated verdict recording
ResultReplaced a large share of manual regression passes; every case triaged into product bug, test drift, or infrastructure block
Root-caused a platform-wide test authentication failure
WhatTraced a long-standing "access denied" failure that was silently blocking a large share of the iOS automated suite — through the account-provisioning service, application-identifier mapping, and token scoping, down to a single incorrect constant
HowToken decoding and diagnostic scripting, API tracing, cross-platform code-path comparison, and verification of the build-and-pin deployment model to prove exactly when the fix went live
ResultRestored blocked coverage — and the same analysis disproved a reported quality regression as sampling variance rather than a real product change
Test-account provisioning infrastructure
WhatExtended the test-account provisioning service to support feature-flag enrollment by default, minor/youth account pairs, non-interactive service-context creation, and propagation-aware waits
HowPython service work, feature-flag enrollment APIs, service-identity credential handling, retry and propagation logic to remove flakiness
ResultUnblocked whole categories of previously untestable coverage — age-restricted flows, parental-control flows, and onboarding paths
AI-Native QA transition
Manual → AI-native testing: full transition
WhatDrove the org-wide shift from fully manual offshore testing to an AI-native, automated-first QA workflow — piloting early tooling, bridging to automation, and building the UX coverage layer that closed the remaining gaps
Scale6–7 months pilot to near-full adoption; 2 apps with full feature suite; daily automated runs replacing weekly manual cycles
ResultOffshore team replaced by automation; weekly → daily test cadence; VP-level recognition for change-based testing methodology
01
Change-based testing agentPiloted AI agent that extracted recent diffs and surfaced relevant exploratory tests. Onboarded and coached offshore team, owned weekly VP-facing reports.
02
Manual → automated bridgeRan parallel manual and automated testing. Identified gaps between outputs, synced weekly with tooling team to inform automation changes. ~1.5 months to convergence.
03
Full automation liveAutomated test passes replaced offshore manual testing entirely. Daily runs replaced weekly cycles.
04
UX gap coverage (ongoing)Built multi-bot pipeline: review automated runs → identify UX issues → dedup against existing bugs → simulator-test net-new issues → package for human review. 2 apps, full feature suite.
Quality operations
Golden Path user journey program
WhatLed partnership with internal beta team, UXR, and Product Design to run structured user journey testing across VR features. Coached non-QA participants on what a useful bug report looks like, then handled the reports they filed
Scale2,600+ survey responses and nearly 400 bugs filed by non-QA testers in a single quarter
ResultTurned people who had never filed a bug into a reliable quality signal, and scaled feedback well beyond the QA team
VR Engagement oncall and release readiness
WhatPrimary QA oncall and release gatekeeper for the VR companion app on iOS and Android — triaging incoming defects, verifying launch-blocking and highest-severity fixes, rerouting mistriaged reports, and signing off releases
Scale6,000+ distinct defects worked across six years and roughly 180 releases — 900+ personally owned to closure, 850+ filed, and 4,500+ triaged, reproduced, or fix-verified for other engineers. 2,300+ were high-severity or emergency priority
Result20% reduction in mis-triaged issues; 40% reduction in production bugs in Meta Horizon Mobile
AI agent development
Agentic workflows
Test automation
Test infrastructure
Root-cause analysis
Release readiness
Cross-platform QA
iOS + Android