QA Engineer for Web Live Stream

Looking for a partner who can independently own a full release cycle alongside the current owner. What this role does day-to-day - End-to-end testing of every release across core features (live streaming / subtitles / recording / cross-platform compatibility, etc.) - Cross-stack diagnosis across Mongo / Redis / Elastic logs / multi-worker pipelines - Proactive coordination across teams (BE / FE / DevOps / upstream services) on test prerequisites, deployment, and release readiness - Producing structured deliverables: test cases / validation reports / bug docs - Actively auditing cross-service version dependencies, cross-environment deployment consistency, and release blockers Baseline technical requirements We don't expect mastery — just enough to operate without friction: Domain depth：At least 2+ years of focused experience in one of: live streaming / video / media streaming / real-time communication / mobile apps Infrastructure：Comfortable with Mongo / Redis commands; able to write KQL queries in Elastic / Kibana CI/CD：Has worked through full deploy pipelines; able to coordinate deployment / rollback with DevOps Debugging / network capture：Comfortable with browser DevTools Network panel; able to infer backend state from API responses Soft skill requirements (core, ordered by importance) 1. Has owned the full "requirement → testing → release" cycle in a structured process At least 1+ year of experience owning complete release cycles, with solid grasp of standard QA process: requirement review / pre-test entry criteria / case review / regression / release readiness / post-release monitoring. 2. Precision in bug triage — knowing that "an issue ≠ a bug" Before filing a bug, the candidate proactively gathers evidence to demonstrate ~90% confidence that it is in fact a bug: Screenshots and screen recordings are baseline, not the goal For non-reliably-reproducible issues, the key skill is identifying the minimal-path repro steps — this is what we actually look for Knows the typical diagnostic axes (environment differences / caching / data state / timing / upstream service versions etc.) Does not file "environment flakiness" or "my own non-standard operation" as bugs 3. Treats PRDs as objects to be tested A product spec itself is something a QA must "test": Understanding the business logic is the baseline, not the goal Proactively surfaces gaps in the PRD: logical holes / edge cases not considered / incorrect premises / implicit assumptions Raises boundary / failure-path / cross-module conflict concerns during requirement review, not after dev implementation Does not accept the working pattern of "wait until dev finishes, then find issues" 4. Multi-dimensional test case design Combines business module × dev implementation × test methodology to design cases covering at minimum: UI checks API logic verification (including failure paths) Data-layer verification (DB state / cache consistency) User-perspective scenario & flow tests Money safety / information security focused cases Performance / capacity related cases Cross-environment / cross-service version dependency audits 5. Multi-party coordination When blocked, knows who to ping (not indiscriminate @-everyone) In multi-party negotiations, can hold the QA position while proposing options acceptable and actionable to all sides Primary goal is steady progress toward release, not "find the most bugs" or "block the release" 6. AI-augmented working style This is a must, not a nice-to-have: BE / FE / DevOps / QA across this project already operate in AI-augmented mode; significant daily work happens inside AI tools The candidate must accept AI as a daily working partner The candidate must actively use AI to amplify their own output: drafting docs / code review / debugging / evidence consolidation / knowledge sedimentation Comfortable exposing their working process to AI; able to collaborate with AI to produce higher-quality output than working solo Has judgment about the boundary of what to expose to AI (what is appropriate, what is not) Nice-to-haves (not required) Hands-on testing experience with live streaming / video pipelines (HLS / VTT / RTMP) Exposure to testing external dependencies like LiveKit / Mux / PubNub A track record of producing structured retrospective docs and similar deliverables History of driving cross-team protocol improvements (e.g. surfacing cross-service dependency blind spots, advocating for an upstream hotfix split, etc.) Prior experience as staff / tech-lead QA What it will look like working together During onboarding: I will walk you through 2-3 complete iteration cycles starting from simpler requirements. This gives you time to internalize our workflow, working patterns, and quality bar — and gives both of us a window to evaluate whether the fit works We will jointly own each release cycle — not split by module, but truly co-owned across versions The current owner is actively iterating their own QA methodology; you are expected to push back, not just absorb Testing periods involve high-density cross-team communication — you will interact directly with BE / FE / DevOps / Product / upstream service teams Most daily work is conducted with AI collaboration: doc drafting / code review / test plan design